diff --git a/cts/cts-fencing.in b/cts/cts-fencing.in index 421e2f6a57..fc4598a648 100644 --- a/cts/cts-fencing.in +++ b/cts/cts-fencing.in @@ -1,1669 +1,1669 @@ #!@PYTHON@ """ Regression tests for Pacemaker's fencer """ __copyright__ = "Copyright 2012-2022 the Pacemaker project contributors" __license__ = "GNU General Public License version 2 or later (GPLv2+) WITHOUT ANY WARRANTY" import io import os import psutil import re import sys import subprocess import shlex import time import tempfile import signal # Where to find test binaries # Prefer the source tree if available BUILD_DIR = "@abs_top_builddir@" SCHEMA_DIR = "@CRM_SCHEMA_DIRECTORY@" COROSYNC_CONF = "@PCMK__COROSYNC_CONF@" TEST_DIR = sys.path[0] AUTOGEN_COROSYNC_TEMPLATE = """ totem { version: 2 cluster_name: cts-fencing crypto_cipher: none crypto_hash: none transport: udp } nodelist { node { nodeid: 1 name: %s ring0_addr: 127.0.0.1 } } logging { debug: off to_syslog: no to_stderr: no to_logfile: yes logfile: %s } """ # These values must be kept in sync with include/crm/crm.h class CrmExit(object): OK = 0 ERROR = 1 INVALID_PARAM = 2 UNIMPLEMENT_FEATURE = 3 INSUFFICIENT_PRIV = 4 NOT_INSTALLED = 5 NOT_CONFIGURED = 6 NOT_RUNNING = 7 USAGE = 64 DATAERR = 65 NOINPUT = 66 NOUSER = 67 NOHOST = 68 UNAVAILABLE = 69 SOFTWARE = 70 OSERR = 71 OSFILE = 72 CANTCREAT = 73 IOERR = 74 TEMPFAIL = 75 PROTOCOL = 76 NOPERM = 77 CONFIG = 78 FATAL = 100 PANIC = 101 DISCONNECT = 102 SOLO = 103 DIGEST = 104 NOSUCH = 105 QUORUM = 106 UNSAFE = 107 EXISTS = 108 MULTIPLE = 109 OLD = 110 TIMEOUT = 124 MAX = 255 def update_path(): """ Set the PATH environment variable appropriately for the tests """ new_path = os.environ['PATH'] if os.path.exists("%s/cts-fencing.in" % TEST_DIR): print("Running tests from the source tree: %s (%s)" % (BUILD_DIR, TEST_DIR)) # For pacemaker-fenced and cts-fence-helper new_path = "%s/daemons/fenced:%s" % (BUILD_DIR, new_path) new_path = "%s/tools:%s" % (BUILD_DIR, new_path) # For stonith_admin new_path = "%s/cts/support:%s" % (BUILD_DIR, new_path) # For cts-support else: print("Running tests from the install tree: @CRM_DAEMON_DIR@ (not %s)" % TEST_DIR) # For pacemaker-fenced, cts-fence-helper, and cts-support new_path = "@CRM_DAEMON_DIR@:%s" % (new_path) print('Using PATH="{}"'.format(new_path)) os.environ['PATH'] = new_path def find_validator(rng_file): if os.access("/usr/bin/xmllint", os.X_OK): if rng_file == None: return ["xmllint", "-"] else: return ["xmllint", "--relaxng", rng_file, "-"] else: return None def rng_directory(): if "PCMK_schema_directory" in os.environ: return os.environ["PCMK_schema_directory"] elif os.path.exists("%s/cts-fencing.in" % TEST_DIR): return "xml" else: return SCHEMA_DIR def pipe_communicate(pipes, check_stderr=False, stdin=None): """ Get text output from pipes """ if stdin is not None: pipe_outputs = pipes.communicate(input=stdin.encode()) else: pipe_outputs = pipes.communicate() output = pipe_outputs[0].decode(sys.stdout.encoding) if check_stderr: output = output + pipe_outputs[1].decode(sys.stderr.encoding) return output def output_from_command(command): """ Execute command and return its standard output """ test = subprocess.Popen(shlex.split(command), stdout=subprocess.PIPE) test.wait() return pipe_communicate(test).split("\n") def localname(): """ Return the uname of the local host """ our_uname = output_from_command("uname -n") if our_uname: our_uname = our_uname[0] else: our_uname = "localhost" return our_uname def killall(process_names=[], terminate=False): """ Kill all instances of every process in a list """ if not process_names: return procs = [] for proc in psutil.process_iter(["name"]): if proc.info["name"] in process_names: procs.append(proc) if terminate: for proc in procs: proc.terminate() gone, alive = psutil.wait_procs(procs, timeout=3) procs = alive for proc in procs: proc.kill() def is_proc_running(process_name): """ Check whether a process with a given name is running """ for proc in psutil.process_iter(["name"]): if proc.info["name"] == process_name: return True return False def exit_if_proc_running(process_name): """ Exit with error if a given process is running """ if is_proc_running(process_name): print("Error: %s is already running!" % process_name) print("Run %s only when the cluster is stopped." % sys.argv[0]) sys.exit(CrmExit.ERROR) class TestError(Exception): """ Base class for exceptions in this module """ pass class ExitCodeError(TestError): """ Exception raised when command exit status is unexpected """ def __init__(self, exit_code): self.exit_code = exit_code def __str__(self): return repr(self.exit_code) class OutputNotFoundError(TestError): """ Exception raised when command output does not contain wanted string """ def __init__(self, output): self.output = output def __str__(self): return repr(self.output) class OutputFoundError(TestError): """ Exception raised when command output contains unwanted string """ def __init__(self, output): self.output = output def __str__(self): return repr(self.output) class XmlValidationError(TestError): """ Exception raised when xmllint fails """ def __init__(self, output): self.output = output def __str__(self): return repr(self.output) class Test(object): """ Executor for a single test """ def __init__(self, name, description, verbose=0, with_cpg=0, timeout=2, force_wait=0, logdir="/tmp"): self.name = name self.description = description self.cmds = [] self.verbose = verbose self.timeout = timeout self.force_wait = force_wait self.logpath = os.path.join(logdir, "pacemaker-fenced.log") self.result_txt = "" self.cmd_tool_output = "" self.result_exitcode = CrmExit.OK if with_cpg: self.stonith_options = "-c" self.enable_corosync = 1 else: self.stonith_options = "-s" self.enable_corosync = 0 self.stonith_process = None self.stonith_output = "" self.stonith_patterns = [] self.negative_stonith_patterns = [] self.executed = 0 def __new_cmd(self, cmd, args, exitcode, stdout_match="", no_wait=0, stdout_negative_match="", kill=None, validate=True, check_rng=True, check_stderr=True): """ Add a command to be executed as part of this test """ self.cmds.append( { "cmd" : cmd, "kill" : kill, "args" : args, "expected_exitcode" : exitcode, "stdout_match" : stdout_match, "stdout_negative_match" : stdout_negative_match, "no_wait" : no_wait, "validate" : validate, "check_rng" : check_rng, "check_stderr" : check_stderr, } ) def start_environment(self): """ Prepare the host for executing a test """ # Make sure we are in full control killall(["pacemakerd", "pacemaker-fenced"]) if self.verbose: self.stonith_options = self.stonith_options + " -V" print("Starting pacemaker-fenced with %s" % self.stonith_options) if os.path.exists(self.logpath): os.remove(self.logpath) cmd = "pacemaker-fenced %s -l %s" % (self.stonith_options, self.logpath) self.stonith_process = subprocess.Popen(shlex.split(cmd)) logfile = None init_time = time.time() update_time = init_time while True: time.sleep(0.1) if self.force_wait == 0 and logfile == None \ and os.path.exists(self.logpath): logfile = io.open(self.logpath, 'rt', encoding = "ISO-8859-1") if self.force_wait == 0 and logfile != None: for line in logfile.readlines(): if "successfully started" in line: return now = time.time() if self.timeout > 0 and (now - init_time) >= self.timeout: if self.force_wait == 0: print("\tDaemon pacemaker-fenced doesn't seem to have been initialized within %fs." "\n\tConsider specifying a longer '--timeout' value." %(self.timeout)) return if self.verbose and (now - update_time) >= 5: print("Waiting for pacemaker-fenced to be initialized: %fs ..." %(now - init_time)) update_time = now def clean_environment(self): """ Clean up the host after executing a test """ if self.stonith_process: if self.stonith_process.poll() == None: self.stonith_process.terminate() self.stonith_process.wait() else: return_code = { getattr(signal, _signame): _signame for _signame in dir(signal) if _signame.startswith('SIG') and not _signame.startswith("SIG_") }.get(-self.stonith_process.returncode, "RET=%d" % (self.stonith_process.returncode)) msg = "FAILURE - '%s' failed. pacemaker-fenced abnormally exited during test (%s)." self.result_txt = msg % (self.name, return_code) self.result_exitcode = CrmExit.ERROR self.stonith_output = "" self.stonith_process = None # the default for utf-8 encoding would error out if e.g. memory corruption # makes fenced output any kind of 8 bit value - while still interesting # for debugging and we'd still like the regression-test to go over the # full set of test-cases logfile = io.open(self.logpath, 'rt', encoding = "ISO-8859-1") for line in logfile.readlines(): self.stonith_output = self.stonith_output + line if self.verbose: print("Daemon Output Start") print(self.stonith_output) print("Daemon Output End") def add_stonith_log_pattern(self, pattern): """ Add a log pattern to expect from this test """ self.stonith_patterns.append(pattern) def add_stonith_neg_log_pattern(self, pattern): """ Add a log pattern that should not occur with this test """ self.negative_stonith_patterns.append(pattern) def add_cmd(self, cmd, args, validate=True, check_rng=True, check_stderr=True): """ Add a simple command to be executed as part of this test """ self.__new_cmd(cmd, args, CrmExit.OK, "", validate=validate, check_rng=check_rng, check_stderr=check_stderr) def add_cmd_no_wait(self, cmd, args): """ Add a simple command to be executed (without waiting) as part of this test """ self.__new_cmd(cmd, args, CrmExit.OK, "", 1) def add_cmd_check_stdout(self, cmd, args, match, no_match=""): """ Add a simple command with expected output to be executed as part of this test """ self.__new_cmd(cmd, args, CrmExit.OK, match, 0, no_match) def add_expected_fail_cmd(self, cmd, args, exitcode=CrmExit.ERROR): """ Add a command to be executed as part of this test and expected to fail """ self.__new_cmd(cmd, args, exitcode, "") def get_exitcode(self): """ Return the exit status of the last test execution """ return self.result_exitcode def print_result(self, filler): """ Print the result of the last test execution """ print("%s%s" % (filler, self.result_txt)) def run_cmd(self, args): """ Execute a command as part of this test """ cmd = shlex.split(args['args']) cmd.insert(0, args['cmd']) if self.verbose: print("\n\nRunning: "+" ".join(cmd)) test = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE) if args['kill']: if self.verbose: print("Also running: "+args['kill']) subprocess.Popen(shlex.split(args['kill'])) if args['no_wait'] == 0: test.wait() else: return CrmExit.OK output = pipe_communicate(test, check_stderr=args['check_stderr']) if self.verbose: print(output) if test.returncode != args['expected_exitcode']: raise ExitCodeError(test.returncode) if (args['stdout_match'] != "" and re.search(args['stdout_match'], output) is None): raise OutputNotFoundError(output) if (args['stdout_negative_match'] != "" and re.search(args['stdout_negative_match'], output) is not None): raise OutputFoundError(output) if args['validate']: if args['check_rng']: rng_file = rng_directory() + "/api/api-result.rng" else: rng_file = None cmd = find_validator(rng_file) if not cmd: return if self.verbose: print("\nRunning: "+" ".join(cmd)) validator = subprocess.Popen(cmd, stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE) output = pipe_communicate(validator, check_stderr=True, stdin=output) if self.verbose: print(output) if validator.returncode != 0: raise XmlValidationError(output) def count_negative_matches(self, outline): """ Return 1 if a line matches patterns that shouldn't have occurred """ count = 0 for line in self.negative_stonith_patterns: if outline.count(line): count = 1 if self.verbose: print("This pattern should not have matched = '%s" % (line)) return count def match_stonith_patterns(self): """ Check test output for expected patterns """ negative_matches = 0 cur = 0 pats = self.stonith_patterns total_patterns = len(self.stonith_patterns) if len(self.stonith_patterns) == 0 and len(self.negative_stonith_patterns) == 0: return for line in self.stonith_output.split("\n"): negative_matches = negative_matches + self.count_negative_matches(line) if len(pats) == 0: continue cur = -1 for pat in pats: cur = cur + 1 if line.count(pats[cur]): del pats[cur] break if len(pats) > 0 or negative_matches: if self.verbose: for pat in pats: print("Pattern Not Matched = '%s'" % pat) msg = "FAILURE - '%s' failed. %d patterns out of %d not matched. %d negative matches." self.result_txt = msg % (self.name, len(pats), total_patterns, negative_matches) self.result_exitcode = CrmExit.ERROR def set_error(self, step, cmd): """ Record failure of this test """ msg = "FAILURE - '%s' failed at step %d. Command: %s %s" self.result_txt = msg % (self.name, step, cmd['cmd'], cmd['args']) self.result_exitcode = CrmExit.ERROR def run(self): """ Execute this test. """ res = 0 i = 1 self.start_environment() if self.verbose: print("\n--- START TEST - %s" % self.name) self.result_txt = "SUCCESS - '%s'" % (self.name) self.result_exitcode = CrmExit.OK for cmd in self.cmds: try: self.run_cmd(cmd) except ExitCodeError as e: print("Step %d FAILED - command returned %s, expected %d" % (i, e, cmd['expected_exitcode'])) self.set_error(i, cmd); break except OutputNotFoundError as e: print("Step %d FAILED - '%s' was not found in command output: %s" % (i, cmd['stdout_match'], e)) self.set_error(i, cmd); break except OutputFoundError as e: print("Step %d FAILED - '%s' was found in command output: %s" % (i, cmd['stdout_negative_match'], e)) self.set_error(i, cmd); break except XmlValidationError as e: print("Step %d FAILED - xmllint failed: %s" % (i, e)) self.set_error(i, cmd); break if self.verbose: print("Step %d SUCCESS" % (i)) i = i + 1 self.clean_environment() if self.result_exitcode == CrmExit.OK: self.match_stonith_patterns() print(self.result_txt) if self.verbose: print("--- END TEST - %s\n" % self.name) self.executed = 1 return res class Tests(object): """ Collection of all fencing regression tests """ def __init__(self, verbose=0, timeout=2, force_wait=0, logdir="/tmp"): self.tests = [] self.verbose = verbose self.timeout = timeout self.force_wait = force_wait self.logdir = logdir self.autogen_corosync_cfg = not os.path.exists(COROSYNC_CONF) def new_test(self, name, description, with_cpg=0): """ Create a named test """ test = Test(name, description, self.verbose, with_cpg, self.timeout, self.force_wait, self.logdir) self.tests.append(test) return test def print_list(self): """ List all registered tests """ print("\n==== %d TESTS FOUND ====" % (len(self.tests))) print("%35s - %s" % ("TEST NAME", "TEST DESCRIPTION")) print("%35s - %s" % ("--------------------", "--------------------")) for test in self.tests: print("%35s - %s" % (test.name, test.description)) print("==== END OF LIST ====\n") def start_corosync(self): """ Start the corosync process """ if self.verbose: print("Starting corosync") test = subprocess.Popen("corosync", stdout=subprocess.PIPE) test.wait() time.sleep(10) def run_single(self, name): """ Run a single named test """ for test in self.tests: if test.name == name: test.run() break def run_tests_matching(self, pattern): """ Run all tests whose name matches a pattern """ for test in self.tests: if test.name.count(pattern) != 0: test.run() def run_cpg_only(self): """ Run all corosync-enabled tests """ for test in self.tests: if test.enable_corosync: test.run() def run_no_cpg(self): """ Run all standalone tests """ for test in self.tests: if not test.enable_corosync: test.run() def run_tests(self): """ Run all tests """ for test in self.tests: test.run() def exit(self): """ Exit (with error status code if any test failed) """ for test in self.tests: if test.executed == 0: continue if test.get_exitcode() != CrmExit.OK: sys.exit(CrmExit.ERROR) sys.exit(CrmExit.OK) def print_results(self): """ Print summary of results of executed tests """ failures = 0 success = 0 print("\n\n======= FINAL RESULTS ==========") print("\n--- FAILURE RESULTS:") for test in self.tests: if test.executed == 0: continue if test.get_exitcode() != CrmExit.OK: failures = failures + 1 test.print_result(" ") else: success = success + 1 if failures == 0: print(" None") print("\n--- TOTALS\n Pass:%d\n Fail:%d\n" % (success, failures)) def build_api_sanity_tests(self): """ Register tests to verify basic API usage """ verbose_arg = "" if self.verbose: verbose_arg = "-V" test = self.new_test("standalone_low_level_api_test", "Sanity test client api in standalone mode.") test.add_cmd("cts-fence-helper", "-t %s" % (verbose_arg), validate=False) test = self.new_test("cpg_low_level_api_test", "Sanity test client api using mainloop and cpg.", 1) test.add_cmd("cts-fence-helper", "-m %s" % (verbose_arg), validate=False) def build_custom_timeout_tests(self): """ Register tests to verify custom timeout usage """ # custom timeout without topology test = self.new_test("cpg_custom_timeout_1", "Verify per device timeouts work as expected without using topology.", 1) test.add_cmd('stonith_admin', '--output-as=xml -R false1 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=node1 node2 node3"') test.add_cmd('stonith_admin', '--output-as=xml -R true1 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node3" -o "pcmk_off_timeout=1"') test.add_cmd('stonith_admin', '--output-as=xml -R false2 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=node3" -o "pcmk_off_timeout=4"') test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 5") # timeout is 5+1+4 = 10 test.add_stonith_log_pattern("Total timeout set to 12") # custom timeout _WITH_ topology test = self.new_test("cpg_custom_timeout_2", "Verify per device timeouts work as expected _WITH_ topology.", 1) test.add_cmd('stonith_admin', '--output-as=xml -R false1 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=node1 node2 node3"') test.add_cmd('stonith_admin', '--output-as=xml -R true1 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node3" -o "pcmk_off_timeout=1"') test.add_cmd('stonith_admin', '--output-as=xml -R false2 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=node3" -o "pcmk_off_timeout=4000"') test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 3 -v false2") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 5") # timeout is 5+1+4000 = 4006 test.add_stonith_log_pattern("Total timeout set to 4807") def build_fence_merge_tests(self): """ Register tests to verify when fence operations should be merged """ ### Simple test that overlapping fencing operations get merged test = self.new_test("cpg_custom_merge_single", "Verify overlapping identical fencing operations are merged, no fencing levels used.", 1) test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node3\" ") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3\"") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 10") ### one merger will happen test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") ### the pattern below signifies that both the original and duplicate operation completed test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") ### Test that multiple mergers occur test = self.new_test("cpg_custom_merge_multiple", "Verify multiple overlapping identical fencing operations are merged", 1) test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"delay=2\" -o \"pcmk_host_list=node3\" ") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3\"") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 10") ### 4 mergers should occur test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") ### the pattern below signifies that both the original and duplicate operation completed test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") ### Test that multiple mergers occur with topologies used test = self.new_test("cpg_custom_merge_with_topology", "Verify multiple overlapping identical fencing operations are merged with fencing levels.", 1) test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node3\" ") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true1") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 10") ### 4 mergers should occur test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") test.add_stonith_log_pattern("Merging fencing action 'off' targeting node3 originating from client") ### the pattern below signifies that both the original and duplicate operation completed test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") test.add_stonith_log_pattern("Operation 'off' targeting node3 by ") def build_fence_no_merge_tests(self): """ Register tests to verify when fence operations should not be merged """ test = self.new_test("cpg_custom_no_merge", "Verify differing fencing operations are not merged", 1) test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3 node2\"") test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node3 node2\" ") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3 node2\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true1") test.add_cmd_no_wait("stonith_admin", "--output-as=xml -F node2 -t 10") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 10") test.add_stonith_neg_log_pattern("Merging fencing action 'off' targeting node3 originating from client") def build_standalone_tests(self): """ Register a grab bag of tests that can be executed in standalone or corosync mode """ test_types = [ { "prefix" : "standalone", "use_cpg" : 0, }, { "prefix" : "cpg", "use_cpg" : 1, }, ] # test what happens when all devices timeout for test_type in test_types: test = self.new_test("%s_fence_multi_device_failure" % test_type["prefix"], "Verify that all devices timeout, a fencing failure is returned.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false3 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") if test_type["use_cpg"] == 1: test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -F node3 -t 2", CrmExit.TIMEOUT) test.add_stonith_log_pattern("Total timeout set to 7") else: test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -F node3 -t 2", CrmExit.ERROR) test.add_stonith_log_pattern("targeting node3 using false1 returned ") test.add_stonith_log_pattern("targeting node3 using false2 returned ") test.add_stonith_log_pattern("targeting node3 using false3 returned ") # test what happens when multiple devices can fence a node, but the first device fails. for test_type in test_types: test = self.new_test("%s_fence_device_failure_rollover" % test_type["prefix"], "Verify that when one fence device fails for a node, the others are tried.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 5") if test_type["use_cpg"] == 1: test.add_stonith_log_pattern("Total timeout set to 18") # test what happens when we try to use a missing fence-agent. for test_type in test_types: test = self.new_test("%s_fence_missing_agent" % test_type["prefix"], "Verify proper error-handling when using a non-existent fence-agent.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_missing -o \"mode=pass\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node2\"") test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -F node3 -t 5", CrmExit.ERROR) test.add_cmd("stonith_admin", "--output-as=xml -F node2 -t 5") # simple topology test for one device for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_topology_simple" % test_type["prefix"], "Verify all fencing devices at a level are used.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v true") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 5") test.add_stonith_log_pattern("Total timeout set to 6") test.add_stonith_log_pattern("targeting node3 using true returned 0") # add topology, delete topology, verify fencing still works for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_topology_add_remove" % test_type["prefix"], "Verify fencing occurrs after all topology levels are removed", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v true") test.add_cmd("stonith_admin", "--output-as=xml -d node3 -i 1") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 5") test.add_stonith_log_pattern("Total timeout set to 6") test.add_stonith_log_pattern("targeting node3 using true returned 0") # test what happens when the first fencing level has multiple devices. for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_topology_device_fails" % test_type["prefix"], "Verify if one device in a level fails, the other is tried.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R false -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 20") test.add_stonith_log_pattern("Total timeout set to 48") test.add_stonith_log_pattern("targeting node3 using false returned 1") test.add_stonith_log_pattern("targeting node3 using true returned 0") # test what happens when the first fencing level fails. for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_topology_multi_level_fails" % test_type["prefix"], "Verify if one level fails, the next leve is tried.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true3 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true4 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v true1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v false2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 3 -v true3") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 3 -v true4") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 3") test.add_stonith_log_pattern("Total timeout set to 21") test.add_stonith_log_pattern("targeting node3 using false1 returned 1") test.add_stonith_log_pattern("targeting node3 using false2 returned 1") test.add_stonith_log_pattern("targeting node3 using true3 returned 0") test.add_stonith_log_pattern("targeting node3 using true4 returned 0") # test what happens when the first fencing level had devices that no one has registered for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_topology_missing_devices" % test_type["prefix"], "Verify topology can continue with missing devices.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true3 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true4 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v true1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v false2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 3 -v true3") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 3 -v true4") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 5") # Test what happens if multiple fencing levels are defined, and then the first one is removed. for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_topology_level_removal" % test_type["prefix"], "Verify level removal works.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true3 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true4 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false2 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v true1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v false2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 3 -v true3") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 3 -v true4") # Now remove level 2, verify none of the devices in level two are hit. test.add_cmd("stonith_admin", "--output-as=xml -d node3 -i 2") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 20") test.add_stonith_log_pattern("Total timeout set to 96") test.add_stonith_log_pattern("targeting node3 using false1 returned 1") test.add_stonith_neg_log_pattern("targeting node3 using false2 returned ") test.add_stonith_log_pattern("targeting node3 using true3 returned 0") test.add_stonith_log_pattern("targeting node3 using true4 returned 0") # Test targeting a topology level by node name pattern. for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_topology_level_pattern" % test_type["prefix"], "Verify targeting topology by node name pattern works.", test_type["use_cpg"]) test.add_cmd("stonith_admin", """--output-as=xml -R true -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node1 node2 node3" """) test.add_cmd("stonith_admin", """--output-as=xml -r '@node.*' -i 1 -v true""") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 5") test.add_stonith_log_pattern("targeting node3 using true returned 0") # test allowing commas and semicolons as delimiters in pcmk_host_list for test_type in test_types: test = self.new_test("%s_host_list_delimiters" % test_type["prefix"], "Verify commas and semicolons can be used as pcmk_host_list delimiters", test_type["use_cpg"]) test.add_cmd("stonith_admin", """--output-as=xml -R true1 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node1,node2,node3" """) test.add_cmd("stonith_admin", """--output-as=xml -R true2 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=pcmk1;pcmk2;pcmk3" """) test.add_cmd("stonith_admin", "stonith_admin --output-as=xml -F node2 -t 5") test.add_cmd("stonith_admin", "stonith_admin --output-as=xml -F pcmk3 -t 5") test.add_stonith_log_pattern("targeting node2 using true1 returned 0") test.add_stonith_log_pattern("targeting pcmk3 using true2 returned 0") # test the stonith builds the correct list of devices that can fence a node. for test_type in test_types: test = self.new_test("%s_list_devices" % test_type["prefix"], "Verify list of devices that can fence a node is correct", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true3 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd_check_stdout("stonith_admin", "--output-as=xml -l node1 -V", "true2", "true1") test.add_cmd_check_stdout("stonith_admin", "--output-as=xml -l node1 -V", "true3", "true1") # simple test of device monitor for test_type in test_types: test = self.new_test("%s_monitor" % test_type["prefix"], "Verify device is reachable", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -Q true1") test.add_cmd("stonith_admin", "--output-as=xml -Q false1") test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -Q true2", CrmExit.ERROR) # Verify monitor occurs for duration of timeout period on failure for test_type in test_types: test = self.new_test("%s_monitor_timeout" % test_type["prefix"], "Verify monitor uses duration of timeout period given.", test_type["use_cpg"]) test.add_cmd("stonith_admin", '--output-as=xml -R true1 -a fence_dummy -o "mode=fail" -o "monitor_mode=fail" -o "pcmk_host_list=node3"') test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -Q true1 -t 5", CrmExit.ERROR) test.add_stonith_log_pattern("Attempt 2 to execute") # Verify monitor occurs for duration of timeout period on failure, but stops at max retries for test_type in test_types: test = self.new_test("%s_monitor_timeout_max_retries" % test_type["prefix"], "Verify monitor retries until max retry value or timeout is hit.", test_type["use_cpg"]) test.add_cmd("stonith_admin", '--output-as=xml -R true1 -a fence_dummy -o "mode=fail" -o "monitor_mode=fail" -o "pcmk_host_list=node3"') test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -Q true1 -t 15", CrmExit.ERROR) test.add_stonith_log_pattern("Attempted to execute agent fence_dummy (list) the maximum number of times") # simple register test for test_type in test_types: test = self.new_test("%s_register" % test_type["prefix"], "Verify devices can be registered and un-registered", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -Q true1") test.add_cmd("stonith_admin", "--output-as=xml -D true1") test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -Q true1", CrmExit.ERROR) # simple reboot test for test_type in test_types: test = self.new_test("%s_reboot" % test_type["prefix"], "Verify devices can be rebooted", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -B node3 -t 5") test.add_cmd("stonith_admin", "--output-as=xml -D true1") test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -Q true1", CrmExit.ERROR) # test fencing history. for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_fence_history" % test_type["prefix"], "Verify last fencing operation is returned.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node3\"") test.add_cmd("stonith_admin", "--output-as=xml -F node3 -t 5 -V") test.add_cmd_check_stdout("stonith_admin", "--output-as=xml -H node3", 'action="off" target="node3" .* status="success"') # simple test of dynamic list query for test_type in test_types: test = self.new_test("%s_dynamic_list_query" % test_type["prefix"], "Verify dynamic list of fencing devices can be retrieved.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o mode=pass -o mock_dynamic_hosts=fake_port_1") test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o mode=pass -o mock_dynamic_hosts=fake_port_1") test.add_cmd("stonith_admin", "--output-as=xml -R true3 -a fence_dummy -o mode=pass -o mock_dynamic_hosts=fake_port_1") test.add_cmd_check_stdout("stonith_admin", "--output-as=xml -l fake_port_1", 'count="3"') # fence using dynamic list query for test_type in test_types: test = self.new_test("%s_fence_dynamic_list_query" % test_type["prefix"], "Verify dynamic list of fencing devices can be retrieved.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o mode=pass -o mock_dynamic_hosts=fake_port_1") test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o mode=pass -o mock_dynamic_hosts=fake_port_1") test.add_cmd("stonith_admin", "--output-as=xml -R true3 -a fence_dummy -o mode=pass -o mock_dynamic_hosts=fake_port_1") test.add_cmd("stonith_admin", "--output-as=xml -F fake_port_1 -t 5 -V") # simple test of query using status action for test_type in test_types: test = self.new_test("%s_status_query" % test_type["prefix"], "Verify dynamic list of fencing devices can be retrieved.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_check=status\"") test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_check=status\"") test.add_cmd("stonith_admin", "--output-as=xml -R true3 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_check=status\"") test.add_cmd_check_stdout("stonith_admin", "--output-as=xml -l fake_port_1", 'count="3"') # test what happens when no reboot action is advertised for test_type in test_types: test = self.new_test("%s_no_reboot_support" % test_type["prefix"], "Verify reboot action defaults to off when no reboot action is advertised by agent.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy_no_reboot -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -B node1 -t 5 -V") test.add_stonith_log_pattern("does not support reboot") test.add_stonith_log_pattern("using true1 returned 0") # make sure reboot is used when reboot action is advertised for test_type in test_types: test = self.new_test("%s_with_reboot_support" % test_type["prefix"], "Verify reboot action can be used when metadata advertises it.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -B node1 -t 5 -V") test.add_stonith_neg_log_pattern("does not advertise support for 'reboot', performing 'off'") test.add_stonith_log_pattern("using true1 returned 0") # make sure requested fencing delay is applied only for the first device in the first level # make sure static delay from pcmk_delay_base is added for test_type in test_types: if test_type["use_cpg"] == 0: continue test = self.new_test("%s_topology_delay" % test_type["prefix"], "Verify requested fencing delay is applied only for the first device in the first level and pcmk_delay_base is added.", test_type["use_cpg"]) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\" -o \"pcmk_delay_base=1\"") test.add_cmd("stonith_admin", "--output-as=xml -R false1 -a fence_dummy -o \"mode=fail\" -o \"pcmk_host_list=node1 node2 node3\" -o \"pcmk_delay_base=1\"") test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -R true3 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=node1 node2 node3\"") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v true1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true2") test.add_cmd("stonith_admin", "--output-as=xml -r node3 -i 2 -v true3") test.add_cmd("stonith_admin", "--output-as=xml -F node3 --delay 1") test.add_stonith_log_pattern("Delaying 'off' action targeting node3 using true1 for 2s | timeout=120s requested_delay=1s base=1s max=1s") test.add_stonith_log_pattern("Delaying 'off' action targeting node3 using false1 for 1s | timeout=120s requested_delay=0s base=1s max=1s") test.add_stonith_neg_log_pattern("Delaying 'off' action targeting node3 using true2") test.add_stonith_neg_log_pattern("Delaying 'off' action targeting node3 using true3") def build_nodeid_tests(self): """ Register tests that use a corosync node id """ our_uname = localname() ### verify nodeid is supplied when nodeid is in the metadata parameters test = self.new_test("cpg_supply_nodeid", "Verify nodeid is given when fence agent has nodeid as parameter", 1) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=%s\"" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -F %s -t 3" % (our_uname)) test.add_stonith_log_pattern("as nodeid with fence action 'off' targeting %s" % (our_uname)) ### verify nodeid is _NOT_ supplied when nodeid is not in the metadata parameters test = self.new_test("cpg_do_not_supply_nodeid", "Verify nodeid is _NOT_ given when fence agent does not have nodeid as parameter", 1) # use a host name that won't be in corosync.conf test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=regr-test\"") test.add_cmd("stonith_admin", "--output-as=xml -F regr-test -t 3") test.add_stonith_neg_log_pattern("as nodeid with fence action 'off' targeting regr-test") ### verify nodeid use doesn't explode standalone mode test = self.new_test("standalone_do_not_supply_nodeid", "Verify nodeid in metadata parameter list doesn't kill standalone mode", 0) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=%s\"" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -F %s -t 3" % (our_uname)) test.add_stonith_neg_log_pattern("as nodeid with fence action 'off' targeting %s" % (our_uname)) def build_unfence_tests(self): """ Register tests that verify unfencing """ our_uname = localname() ### verify unfencing using automatic unfencing test = self.new_test("cpg_unfence_required_1", "Verify require unfencing on all devices when automatic=true in agent's metadata", 1) test.add_cmd('stonith_admin', '--output-as=xml -R true1 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R true2 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s"' % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -U %s -t 3" % (our_uname)) # both devices should be executed test.add_stonith_log_pattern("using true1 returned 0") test.add_stonith_log_pattern("using true2 returned 0") ### verify unfencing using automatic unfencing fails if any of the required agents fail test = self.new_test("cpg_unfence_required_2", "Verify require unfencing on all devices when automatic=true in agent's metadata", 1) test.add_cmd('stonith_admin', '--output-as=xml -R true1 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R true2 -a fence_dummy_auto_unfence -o "mode=fail" -o "pcmk_host_list=%s"' % (our_uname)) test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -U %s -t 6" % (our_uname), CrmExit.ERROR) ### verify unfencing using automatic devices with topology test = self.new_test("cpg_unfence_required_3", "Verify require unfencing on all devices even when at different topology levels", 1) test.add_cmd('stonith_admin', '--output-as=xml -R true1 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R true2 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 1 -v true1" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 2 -v true2" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -U %s -t 3" % (our_uname)) test.add_stonith_log_pattern("using true1 returned 0") test.add_stonith_log_pattern("using true2 returned 0") ### verify unfencing using automatic devices with topology test = self.new_test("cpg_unfence_required_4", "Verify all required devices are executed even with topology levels fail.", 1) test.add_cmd('stonith_admin', '--output-as=xml -R true1 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R true2 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R true3 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R true4 -a fence_dummy_auto_unfence -o "mode=pass" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R false1 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R false2 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R false3 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd('stonith_admin', '--output-as=xml -R false4 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=%s node3"' % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 1 -v true1" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 1 -v false1" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 2 -v false2" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 2 -v true2" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 2 -v false3" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 2 -v true3" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 3 -v false4" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 4 -v true4" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -U %s -t 3" % (our_uname)) test.add_stonith_log_pattern("using true1 returned 0") test.add_stonith_log_pattern("using true2 returned 0") test.add_stonith_log_pattern("using true3 returned 0") test.add_stonith_log_pattern("using true4 returned 0") def build_unfence_on_target_tests(self): """ Register tests that verify unfencing that runs on the target """ our_uname = localname() ### verify unfencing using on_target device test = self.new_test("cpg_unfence_on_target_1", "Verify unfencing with on_target = true", 1) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=%s\"" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -U %s -t 3" % (our_uname)) test.add_stonith_log_pattern("(on) to be executed on target") ### verify failure of unfencing using on_target device test = self.new_test("cpg_unfence_on_target_2", "Verify failure unfencing with on_target = true", 1) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=%s node_fake_1234\"" % (our_uname)) test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -U node_fake_1234 -t 3", CrmExit.ERROR) test.add_stonith_log_pattern("(on) to be executed on target") ### verify unfencing using on_target device with topology test = self.new_test("cpg_unfence_on_target_3", "Verify unfencing with on_target = true using topology", 1) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=%s node3\"" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=%s node3\"" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 1 -v true1" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r %s -i 2 -v true2" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -U %s -t 3" % (our_uname)) test.add_stonith_log_pattern("(on) to be executed on target") - ### verify unfencing using on_target device with topology fails when victim node doesn't exist + ### verify unfencing using on_target device with topology fails when target node doesn't exist test = self.new_test("cpg_unfence_on_target_4", "Verify unfencing failure with on_target = true using topology", 1) test.add_cmd("stonith_admin", "--output-as=xml -R true1 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=%s node_fake\"" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -R true2 -a fence_dummy -o \"mode=pass\" -o \"pcmk_host_list=%s node_fake\"" % (our_uname)) test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 1 -v true1") test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 2 -v true2") test.add_expected_fail_cmd("stonith_admin", "--output-as=xml -U node_fake -t 3", CrmExit.ERROR) test.add_stonith_log_pattern("(on) to be executed on target") def build_remap_tests(self): """ Register tests that verify remapping of reboots to off-on """ test = self.new_test("cpg_remap_simple", "Verify sequential topology reboot is remapped to all-off-then-all-on", 1) test.add_cmd("stonith_admin", """--output-as=xml -R true1 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node_fake" """ """-o "pcmk_off_timeout=1" -o "pcmk_reboot_timeout=10" """) test.add_cmd("stonith_admin", """--output-as=xml -R true2 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node_fake" """ """-o "pcmk_off_timeout=2" -o "pcmk_reboot_timeout=20" """) test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 1 -v true1 -v true2") test.add_cmd("stonith_admin", "--output-as=xml -B node_fake -t 5") test.add_stonith_log_pattern("Remapping multiple-device reboot targeting node_fake") # timeout should be sum of off timeouts (1+2=3), not reboot timeouts (10+20=30) test.add_stonith_log_pattern("Total timeout set to 3 for peer's fencing targeting node_fake") test.add_stonith_log_pattern("perform 'off' action targeting node_fake using true1") test.add_stonith_log_pattern("perform 'off' action targeting node_fake using true2") test.add_stonith_log_pattern("Remapped 'off' targeting node_fake complete, remapping to 'on'") # fence_dummy sets "on" as an on_target action test.add_stonith_log_pattern("Ignoring true1 'on' failure (no capable peers) targeting node_fake") test.add_stonith_log_pattern("Ignoring true2 'on' failure (no capable peers) targeting node_fake") test.add_stonith_log_pattern("Undoing remap of reboot targeting node_fake") test = self.new_test("cpg_remap_automatic", "Verify remapped topology reboot skips automatic 'on'", 1) test.add_cmd("stonith_admin", """--output-as=xml -R true1 -a fence_dummy_auto_unfence """ """-o "mode=pass" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", """--output-as=xml -R true2 -a fence_dummy_auto_unfence """ """-o "mode=pass" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 1 -v true1 -v true2") test.add_cmd("stonith_admin", "--output-as=xml -B node_fake -t 5") test.add_stonith_log_pattern("Remapping multiple-device reboot targeting node_fake") test.add_stonith_log_pattern("perform 'off' action targeting node_fake using true1") test.add_stonith_log_pattern("perform 'off' action targeting node_fake using true2") test.add_stonith_log_pattern("Remapped 'off' targeting node_fake complete, remapping to 'on'") test.add_stonith_log_pattern("Undoing remap of reboot targeting node_fake") test.add_stonith_neg_log_pattern("perform 'on' action targeting node_fake using") test.add_stonith_neg_log_pattern("'on' failure") test = self.new_test("cpg_remap_complex_1", "Verify remapped topology reboot in second level works if non-remapped first level fails", 1) test.add_cmd("stonith_admin", """--output-as=xml -R false1 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", """--output-as=xml -R true1 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", """--output-as=xml -R true2 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 2 -v true1 -v true2") test.add_cmd("stonith_admin", "--output-as=xml -B node_fake -t 5") test.add_stonith_log_pattern("perform 'reboot' action targeting node_fake using false1") test.add_stonith_log_pattern("Remapping multiple-device reboot targeting node_fake") test.add_stonith_log_pattern("perform 'off' action targeting node_fake using true1") test.add_stonith_log_pattern("perform 'off' action targeting node_fake using true2") test.add_stonith_log_pattern("Remapped 'off' targeting node_fake complete, remapping to 'on'") test.add_stonith_log_pattern("Ignoring true1 'on' failure (no capable peers) targeting node_fake") test.add_stonith_log_pattern("Ignoring true2 'on' failure (no capable peers) targeting node_fake") test.add_stonith_log_pattern("Undoing remap of reboot targeting node_fake") test = self.new_test("cpg_remap_complex_2", "Verify remapped topology reboot failure in second level proceeds to third level", 1) test.add_cmd("stonith_admin", """--output-as=xml -R false1 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", """--output-as=xml -R false2 -a fence_dummy -o "mode=fail" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", """--output-as=xml -R true1 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", """--output-as=xml -R true2 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", """--output-as=xml -R true3 -a fence_dummy -o "mode=pass" -o "pcmk_host_list=node_fake" """) test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 1 -v false1") test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 2 -v true1 -v false2 -v true3") test.add_cmd("stonith_admin", "--output-as=xml -r node_fake -i 3 -v true2") test.add_cmd("stonith_admin", "--output-as=xml -B node_fake -t 5") test.add_stonith_log_pattern("perform 'reboot' action targeting node_fake using false1") test.add_stonith_log_pattern("Remapping multiple-device reboot targeting node_fake") test.add_stonith_log_pattern("perform 'off' action targeting node_fake using true1") test.add_stonith_log_pattern("perform 'off' action targeting node_fake using false2") test.add_stonith_log_pattern("Attempted to execute agent fence_dummy (off) the maximum number of times") test.add_stonith_log_pattern("Undoing remap of reboot targeting node_fake") test.add_stonith_log_pattern("perform 'reboot' action targeting node_fake using true2") test.add_stonith_neg_log_pattern("node_fake with true3") def build_query_tests(self): """ run stonith_admin --metadata for the fence_dummy agent and check command output """ test = self.new_test("get_metadata", "Run stonith_admin --metadata for the fence_dummy agent", 1) test.add_cmd_check_stdout("stonith_admin", "--output-as=xml -a fence_dummy --metadata", ' %s %s """ % (agent(), SHORT_DESC, AGENT_VERSION, OCF_VERSION, LONG_DESC)) for option, dummy in sorted_options(avail_opt): if "shortdesc" in ALL_OPT[option]: print(' ') default = "" default_name_arg = "-" + ALL_OPT[option]["getopt"][:-1] default_name_no_arg = "-" + ALL_OPT[option]["getopt"] if "default" in ALL_OPT[option]: default = 'default="%s"' % str(ALL_OPT[option]["default"]) elif default_name_arg in options: if options[default_name_arg]: try: default = 'default="%s"' % options[default_name_arg] except TypeError: ## @todo/@note: Currently there is no clean way how to handle lists ## we can create a string from it but we can't set it on command line default = 'default="%s"' % str(options[default_name_arg]) elif default_name_no_arg in options: default = 'default="true"' mixed = ALL_OPT[option]["help"] ## split it between option and help text res = re.compile(r"^(.*--\S+)\s+", re.IGNORECASE | re.S).search(mixed) if None != res: mixed = res.group(1) mixed = mixed.replace("<", "<").replace(">", ">") print(' ') if ALL_OPT[option]["getopt"].count(":") > 0: print(' ') else: print(' ') print(' ' + ALL_OPT[option]["shortdesc"] + '') print(' ') print(' \n ') if auto_unfence: attr_name = 'automatic' else: attr_name = 'on_target' print(' ') print(' ') if not no_reboot: print(' ') print(' ') print(' ') print(' ') print(' ') print(' ') print('') def option_longopt(option): """ Return the getopt-compatible long-option name of the given option. """ if ALL_OPT[option]["getopt"].endswith(":"): return ALL_OPT[option]["longopt"] + "=" else: return ALL_OPT[option]["longopt"] def opts_from_command_line(argv, avail_opt): """ Read options from command-line arguments. """ # Prepare list of options for getopt getopt_string = "" longopt_list = [] for k in avail_opt: if k in ALL_OPT: getopt_string += ALL_OPT[k]["getopt"] else: fail_usage("Parse error: unknown option '"+k+"'") if k in ALL_OPT and "longopt" in ALL_OPT[k]: longopt_list.append(option_longopt(k)) try: opt, dummy = getopt.gnu_getopt(argv, getopt_string, longopt_list) except getopt.GetoptError as error: fail_usage("Parse error: " + error.msg) # Transform longopt to short one which are used in fencing agents old_opt = opt opt = {} for old_option in dict(old_opt).keys(): if old_option.startswith("--"): for option in ALL_OPT.keys(): if "longopt" in ALL_OPT[option] and "--" + ALL_OPT[option]["longopt"] == old_option: opt["-" + ALL_OPT[option]["getopt"].rstrip(":")] = dict(old_opt)[old_option] else: opt[old_option] = dict(old_opt)[old_option] # Compatibility Layer (with what? probably not needed for fence_dummy) new_opt = dict(opt) if "-T" in new_opt: new_opt["-o"] = "status" if "-n" in new_opt: new_opt["-m"] = new_opt["-n"] opt = new_opt return opt def opts_from_stdin(avail_opt): """ Read options from standard input. """ opt = {} name = "" for line in sys.stdin.readlines(): line = line.strip() if line.startswith("#") or (len(line) == 0): continue (name, value) = (line + "=").split("=", 1) value = value[:-1] # Compatibility Layer (with what? probably not needed for fence_dummy) if name == "option": name = "action" if name not in avail_opt: print("Parse error: Ignoring unknown option '%s'" % line, file=sys.stderr) continue if ALL_OPT[name]["getopt"].endswith(":"): opt["-"+ALL_OPT[name]["getopt"].rstrip(":")] = value elif value.lower() in ["1", "yes", "on", "true"]: opt["-"+ALL_OPT[name]["getopt"]] = "1" return opt def process_input(avail_opt): """ Set standard environment variables, and parse all options. """ # Set standard environment os.putenv("LANG", "C") os.putenv("LC_ALL", "C") # Read options from command line or standard input if len(sys.argv) > 1: return opts_from_command_line(sys.argv[1:], avail_opt) else: return opts_from_stdin(avail_opt) def atexit_handler(): """ Close stdout on exit. """ try: sys.stdout.close() os.close(1) except IOError: sys.exit("%s failed to close standard output" % agent()) def success_mode(options, option, default_value): """ Return exit code specified by option. """ if option in options: test_value = options[option] else: test_value = default_value if test_value == "pass": exitcode = 0 elif test_value == "fail": exitcode = 1 else: exitcode = random.randint(0, 1) return exitcode def write_options(options): """ Write out all options to debug file. """ try: debugfile = io.open(options["-D"], 'at') debugfile.write("### %s ###\n" % (time.strftime("%Y-%m-%d %H:%M:%S"))) for option in sorted(options): debugfile.write("%s=%s\n" % (option, options[option])) debugfile.write("###\n") debugfile.close() except IOError: pass def main(): """ Make it so! """ global auto_unfence global no_reboot # Meta-data can't take parameters, so we simulate different meta-data # behavior based on the executable name (which can be a symbolic link). if (sys.argv[0].endswith("_auto_unfence")): auto_unfence = True elif (sys.argv[0].endswith("_no_reboot")): no_reboot = True device_opt = ALL_OPT.keys() ## Defaults for fence agent atexit.register(atexit_handler) options = process_input(device_opt) options["device_opt"] = device_opt show_docs(options) if "-o" in options: action = options["-o"] else: action = "reboot" # dump input to file if "-D" in options and action != "validate-all": write_options(options) if "-f" in options and action != "validate-all": val = int(options["-f"]) print("delay sleep for %d seconds" % val, file=sys.stderr) time.sleep(val) # random sleep for testing if "-R" in options and action != "validate-all": val = int(options["-R"]) ran = random.randint(1, val) print("random sleep for %d seconds" % ran, file=sys.stderr) time.sleep(ran) if action == "monitor": if "-d" in options: time.sleep(int(options["-d"])) exitcode = success_mode(options, "-m", "pass") elif action == "list": print("fence_dummy action (list) called", file=sys.stderr) if "-H" in options: print(options["-H"]) exitcode = 0 else: print("dynamic hostlist requires mock_dynamic_hosts to be set", file=sys.stderr) exitcode = 1 elif action == "validate-all": if "-f" in options: val = int(options["-f"]) if val > 10: exitcode = 1 else: exitcode = 0 else: exitcode = 1 else: exitcode = success_mode(options, "-M", "random") # Ensure we generate some error output on failure exit. if exitcode == 1: print("simulated %s failure" % action, file=sys.stderr) sys.exit(exitcode) if __name__ == "__main__": main() diff --git a/daemons/fenced/fence_watchdog.in b/daemons/fenced/fence_watchdog.in index eefa7395e1..f43ab879d4 100755 --- a/daemons/fenced/fence_watchdog.in +++ b/daemons/fenced/fence_watchdog.in @@ -1,284 +1,284 @@ #!@PYTHON@ """Dummy watchdog fence agent for providing meta-data for the pacemaker internal agent """ -__copyright__ = "Copyright 2012-2021 the Pacemaker project contributors" +__copyright__ = "Copyright 2012-2022 the Pacemaker project contributors" __license__ = "GNU General Public License version 2 or later (GPLv2+) WITHOUT ANY WARRANTY" import io import os import re import sys import atexit import getopt AGENT_VERSION = "1.0.0" SHORT_DESC = "Dummy watchdog fence agent" LONG_DESC = """fence_watchdog just provides meta-data - actual fencing is done by the pacemaker internal watchdog agent.""" ALL_OPT = { "version" : { "getopt" : "V", "longopt" : "version", "help" : "-V, --version Display version information and exit", "required" : "0", "shortdesc" : "Display version information and exit", "order" : 53 }, "help" : { "getopt" : "h", "longopt" : "help", "help" : "-h, --help Display this help and exit", "required" : "0", "shortdesc" : "Display help and exit", "order" : 54 }, "action" : { "getopt" : "o:", "longopt" : "action", "help" : "-o, --action=[action] Action: metadata", "required" : "1", "shortdesc" : "Fencing Action", "default" : "metadata", "order" : 1 }, "nodename" : { "getopt" : "N:", "longopt" : "nodename", - "help" : "-N, --nodename Node name of fence victim (ignored)", + "help" : "-N, --nodename Node name of fence target (ignored)", "required" : "0", "shortdesc" : "Ignored", "order" : 2 }, "plug" : { "getopt" : "n:", "longopt" : "plug", "help" : "-n, --plug=[id] Physical plug number on device (ignored)", "required" : "1", "shortdesc" : "Ignored", "order" : 4 } } def agent(): """ Return name this file was run as. """ return os.path.basename(sys.argv[0]) def fail_usage(message): """ Print a usage message and exit. """ sys.exit("%s\nPlease use '-h' for usage" % message) def show_docs(options): """ Handle informational options (display info and exit). """ device_opt = options["device_opt"] if "-h" in options: usage(device_opt) sys.exit(0) if "-o" in options and options["-o"].lower() == "metadata": metadata(device_opt, options) sys.exit(0) if "-V" in options: print(AGENT_VERSION) sys.exit(0) def sorted_options(avail_opt): """ Return a list of all options, in their internally specified order. """ sorted_list = [(key, ALL_OPT[key]) for key in avail_opt] sorted_list.sort(key=lambda x: x[1]["order"]) return sorted_list def usage(avail_opt): """ Print a usage message. """ print(LONG_DESC) print() print("Usage:") print("\t" + agent() + " [options]") print("Options:") for dummy, value in sorted_options(avail_opt): if len(value["help"]) != 0: print(" " + value["help"]) def metadata(avail_opt, options): """ Print agent metadata. """ print(""" %s """ % (agent(), SHORT_DESC, LONG_DESC)) for option, dummy in sorted_options(avail_opt): if "shortdesc" in ALL_OPT[option]: print(' ') default = "" default_name_arg = "-" + ALL_OPT[option]["getopt"][:-1] default_name_no_arg = "-" + ALL_OPT[option]["getopt"] if "default" in ALL_OPT[option]: default = 'default="%s"' % str(ALL_OPT[option]["default"]) elif default_name_arg in options: if options[default_name_arg]: try: default = 'default="%s"' % options[default_name_arg] except TypeError: ## @todo/@note: Currently there is no clean way how to handle lists ## we can create a string from it but we can't set it on command line default = 'default="%s"' % str(options[default_name_arg]) elif default_name_no_arg in options: default = 'default="true"' mixed = ALL_OPT[option]["help"] ## split it between option and help text res = re.compile(r"^(.*--\S+)\s+", re.IGNORECASE | re.S).search(mixed) if None != res: mixed = res.group(1) mixed = mixed.replace("<", "<").replace(">", ">") print(' ') if ALL_OPT[option]["getopt"].count(":") > 0: print(' ') else: print(' ') print(' ' + ALL_OPT[option]["shortdesc"] + '') print(' ') print(' \n ') print(' ') print(' ') print(' ') print(' ') print(' ') print(' ') print(' ') print('') def option_longopt(option): """ Return the getopt-compatible long-option name of the given option. """ if ALL_OPT[option]["getopt"].endswith(":"): return ALL_OPT[option]["longopt"] + "=" else: return ALL_OPT[option]["longopt"] def opts_from_command_line(argv, avail_opt): """ Read options from command-line arguments. """ # Prepare list of options for getopt getopt_string = "" longopt_list = [] for k in avail_opt: if k in ALL_OPT: getopt_string += ALL_OPT[k]["getopt"] else: fail_usage("Parse error: unknown option '" + k + "'") if k in ALL_OPT and "longopt" in ALL_OPT[k]: longopt_list.append(option_longopt(k)) try: opt, dummy = getopt.gnu_getopt(argv, getopt_string, longopt_list) except getopt.GetoptError as error: fail_usage("Parse error: " + error.msg) # Transform longopt to short one which are used in fencing agents old_opt = opt opt = {} for old_option in dict(old_opt).keys(): if old_option.startswith("--"): for option in ALL_OPT.keys(): if "longopt" in ALL_OPT[option] and "--" + ALL_OPT[option]["longopt"] == old_option: opt["-" + ALL_OPT[option]["getopt"].rstrip(":")] = dict(old_opt)[old_option] else: opt[old_option] = dict(old_opt)[old_option] return opt def opts_from_stdin(avail_opt): """ Read options from standard input. """ opt = {} name = "" for line in sys.stdin.readlines(): line = line.strip() if line.startswith("#") or (len(line) == 0): continue (name, value) = (line + "=").split("=", 1) value = value[:-1] if name not in avail_opt: print("Parse error: Ignoring unknown option '%s'" % line, file=sys.stderr) continue if ALL_OPT[name]["getopt"].endswith(":"): opt["-"+ALL_OPT[name]["getopt"].rstrip(":")] = value elif value.lower() in ["1", "yes", "on", "true"]: opt["-"+ALL_OPT[name]["getopt"]] = "1" return opt def process_input(avail_opt): """ Set standard environment variables, and parse all options. """ # Set standard environment os.putenv("LANG", "C") os.putenv("LC_ALL", "C") # Read options from command line or standard input if len(sys.argv) > 1: return opts_from_command_line(sys.argv[1:], avail_opt) else: return opts_from_stdin(avail_opt) def atexit_handler(): """ Close stdout on exit. """ try: sys.stdout.close() os.close(1) except IOError: sys.exit("%s failed to close standard output" % agent()) def main(): """ Make it so! """ device_opt = ALL_OPT.keys() ## Defaults for fence agent atexit.register(atexit_handler) options = process_input(device_opt) options["device_opt"] = device_opt show_docs(options) print("Watchdog fencing may be initiated only by the cluster, not this agent.", file=sys.stderr) sys.exit(1) if __name__ == "__main__": main() diff --git a/daemons/fenced/pacemaker-fenced.h b/daemons/fenced/pacemaker-fenced.h index acd14fd927..7992381531 100644 --- a/daemons/fenced/pacemaker-fenced.h +++ b/daemons/fenced/pacemaker-fenced.h @@ -1,290 +1,292 @@ /* * Copyright 2009-2022 the Pacemaker project contributors * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include // uint32_t, uint64_t #include /*! * \internal * \brief Check whether target has already been fenced recently * * \param[in] tolerance Number of seconds to look back in time * \param[in] target Name of node to search for * \param[in] action Action we want to match * * \return TRUE if an equivalent fencing operation took place in the last * \p tolerance seconds, FALSE otherwise */ gboolean stonith_check_fence_tolerance(int tolerance, const char *target, const char *action); typedef struct stonith_device_s { char *id; char *agent; char *namespace; /*! list of actions that must execute on the target node. Used for unfencing */ char *on_target_actions; GList *targets; time_t targets_age; gboolean has_attr_map; - /* should nodeid parameter for victim be included in agent arguments */ + + // Whether target's nodeid should be passed as a parameter to the agent gboolean include_nodeid; + /* whether the cluster should automatically unfence nodes with the device */ gboolean automatic_unfencing; guint priority; uint32_t flags; // Group of enum st_device_flags GHashTable *params; GHashTable *aliases; GList *pending_ops; mainloop_timer_t *timer; crm_trigger_t *work; xmlNode *agent_metadata; /*! A verified device is one that has contacted the * agent successfully to perform a monitor operation */ gboolean verified; gboolean cib_registered; gboolean api_registered; gboolean dirty; } stonith_device_t; /* These values are used to index certain arrays by "phase". Usually an * operation has only one "phase", so phase is always zero. However, some * reboots are remapped to "off" then "on", in which case "reboot" will be * phase 0, "off" will be phase 1 and "on" will be phase 2. */ enum st_remap_phase { st_phase_requested = 0, st_phase_off = 1, st_phase_on = 2, st_phase_max = 3 }; typedef struct remote_fencing_op_s { /* The unique id associated with this operation */ char *id; /*! The node this operation will fence */ char *target; /*! The fencing action to perform on the target. (reboot, on, off) */ char *action; /*! When was the fencing action recorded (seconds since epoch) */ time_t created; /*! Marks if the final notifications have been sent to local stonith clients. */ gboolean notify_sent; /*! The number of query replies received */ guint replies; /*! The number of query replies expected */ guint replies_expected; /*! Does this node own control of this operation */ gboolean owner; /*! After query is complete, This the high level timer that expires the entire operation */ guint op_timer_total; /*! This timer expires the current fencing request. Many fencing * requests may exist in a single operation */ guint op_timer_one; /*! This timer expires the query request sent out to determine * what nodes are contain what devices, and who those devices can fence */ guint query_timer; /*! This is the default timeout to use for each fencing device if no * custom timeout is received in the query. */ gint base_timeout; /*! This is the calculated total timeout an operation can take before * expiring. This is calculated by adding together all the timeout * values associated with the devices this fencing operation may call */ gint total_timeout; /*! Requested fencing delay. * Value -1 means disable any static/random fencing delays. */ int delay; /*! Delegate is the node being asked to perform a fencing action * on behalf of the node that owns the remote operation. Some operations * will involve multiple delegates. This value represents the final delegate * that is used. */ char *delegate; /*! The point at which the remote operation completed */ time_t completed; //! Group of enum stonith_call_options associated with this operation uint32_t call_options; /*! The current state of the remote operation. This indicates * what stage the op is in, query, exec, done, duplicate, failed. */ enum op_state state; /*! The node that owns the remote operation */ char *originator; /*! The local client id that initiated the fencing request */ char *client_id; /*! The client's call_id that initiated the fencing request */ int client_callid; /*! The name of client that initiated the fencing request */ char *client_name; /*! List of the received query results for all the nodes in the cpg group */ GList *query_results; /*! The original request that initiated the remote stonith operation */ xmlNode *request; /*! The current topology level being executed */ guint level; /*! The current operation phase being executed */ enum st_remap_phase phase; /*! Devices with automatic unfencing (always run if "on" requested, never if remapped) */ GList *automatic_list; /*! List of all devices at the currently executing topology level */ GList *devices_list; /*! Current entry in the topology device list */ GList *devices; /*! List of duplicate operations attached to this operation. Once this operation * completes, the duplicate operations will be closed out as well. */ GList *duplicates; /*! The point at which the remote operation completed(nsec) */ long long completed_nsec; /*! The (potentially intermediate) result of the operation */ pcmk__action_result_t result; } remote_fencing_op_t; void fenced_broadcast_op_result(remote_fencing_op_t *op, bool op_merged); // Fencer-specific client flags enum st_client_flags { st_callback_unknown = UINT64_C(0), st_callback_notify_fence = (UINT64_C(1) << 0), st_callback_device_add = (UINT64_C(1) << 2), st_callback_device_del = (UINT64_C(1) << 4), st_callback_notify_history = (UINT64_C(1) << 5), st_callback_notify_history_synced = (UINT64_C(1) << 6) }; // How the user specified the target of a topology level enum fenced_target_by { fenced_target_by_unknown = -1, // Invalid or not yet parsed fenced_target_by_name, // By target name fenced_target_by_pattern, // By a pattern matching target names fenced_target_by_attribute, // By a node attribute/value on target }; /* * Complex fencing requirements are specified via fencing topologies. * A topology consists of levels; each level is a list of fencing devices. * Topologies are stored in a hash table by node name. When a node needs to be * fenced, if it has an entry in the topology table, the levels are tried * sequentially, and the devices in each level are tried sequentially. * Fencing is considered successful as soon as any level succeeds; * a level is considered successful if all its devices succeed. * Essentially, all devices at a given level are "and-ed" and the * levels are "or-ed". * * This structure is used for the topology table entries. * Topology levels start from 1, so levels[0] is unused and always NULL. */ typedef struct stonith_topology_s { enum fenced_target_by kind; // How target was specified /*! Node name regex or attribute name=value for which topology applies */ char *target; char *target_value; char *target_pattern; char *target_attribute; /*! Names of fencing devices at each topology level */ GList *levels[ST_LEVEL_MAX]; } stonith_topology_t; void init_device_list(void); void free_device_list(void); void init_topology_list(void); void free_topology_list(void); void free_stonith_remote_op_list(void); void init_stonith_remote_op_hash_table(GHashTable **table); void free_metadata_cache(void); void fenced_unregister_handlers(void); uint64_t get_stonith_flag(const char *name); void stonith_command(pcmk__client_t *client, uint32_t id, uint32_t flags, xmlNode *op_request, const char *remote_peer); int stonith_device_register(xmlNode *msg, gboolean from_cib); void stonith_device_remove(const char *id, bool from_cib); char *stonith_level_key(xmlNode * msg, int mode); void fenced_register_level(xmlNode *msg, char **desc, pcmk__action_result_t *result); void fenced_unregister_level(xmlNode *msg, char **desc, pcmk__action_result_t *result); stonith_topology_t *find_topology_for_host(const char *host); void do_local_reply(xmlNode *notify_src, pcmk__client_t *client, int call_options); xmlNode *fenced_construct_reply(xmlNode *request, xmlNode *data, pcmk__action_result_t *result); void do_stonith_async_timeout_update(const char *client, const char *call_id, int timeout); void fenced_send_notification(const char *type, const pcmk__action_result_t *result, xmlNode *data); void fenced_send_device_notification(const char *op, const pcmk__action_result_t *result, const char *desc); void fenced_send_level_notification(const char *op, const pcmk__action_result_t *result, const char *desc); remote_fencing_op_t *initiate_remote_stonith_op(pcmk__client_t *client, xmlNode *request, gboolean manual_ack); void fenced_process_fencing_reply(xmlNode *msg); int process_remote_stonith_query(xmlNode * msg); void *create_remote_stonith_op(const char *client, xmlNode * request, gboolean peer); void stonith_fence_history(xmlNode *msg, xmlNode **output, const char *remote_peer, int options); void stonith_fence_history_trim(void); bool fencing_peer_active(crm_node_t *peer); void set_fencing_completed(remote_fencing_op_t * op); int fenced_handle_manual_confirmation(pcmk__client_t *client, xmlNode *msg); gboolean node_has_attr(const char *node, const char *name, const char *value); gboolean node_does_watchdog_fencing(const char *node); static inline void fenced_set_protocol_error(pcmk__action_result_t *result) { pcmk__set_result(result, CRM_EX_PROTOCOL, PCMK_EXEC_INVALID, "Fencer API request missing required information (bug?)"); } extern char *stonith_our_uname; extern gboolean stand_alone; extern GHashTable *device_list; extern GHashTable *topology; extern long stonith_watchdog_timeout_ms; extern GList *stonith_watchdog_targets; extern GHashTable *stonith_remote_op_list; diff --git a/doc/sphinx/Clusters_from_Scratch/fencing.rst b/doc/sphinx/Clusters_from_Scratch/fencing.rst index ae3aba20fe..65537bfdf1 100644 --- a/doc/sphinx/Clusters_from_Scratch/fencing.rst +++ b/doc/sphinx/Clusters_from_Scratch/fencing.rst @@ -1,231 +1,231 @@ .. index:: fencing Configure Fencing ----------------- What is Fencing? ################ Fencing protects your data from being corrupted, and your application from becoming unavailable, due to unintended concurrent access by rogue nodes. Just because a node is unresponsive doesn't mean it has stopped accessing your data. The only way to be 100% sure that your data is safe, is to use fencing to ensure that the node is truly offline before allowing the data to be accessed from another node. Fencing also has a role to play in the event that a clustered service cannot be stopped. In this case, the cluster uses fencing to force the whole node offline, thereby making it safe to start the service elsewhere. Fencing is also known as STONITH, an acronym for "Shoot The Other Node In The Head", since the most popular form of fencing is cutting a host's power. In order to guarantee the safety of your data [#]_, fencing is enabled by default. .. NOTE:: It is possible to tell the cluster not to use fencing, by setting the ``stonith-enabled`` cluster property to false: .. code-block:: console [root@pcmk-1 ~]# pcs property set stonith-enabled=false [root@pcmk-1 ~]# pcs cluster verify --full However, this is completely inappropriate for a production cluster. It tells the cluster to simply pretend that failed nodes are safely powered off. Some vendors will refuse to support clusters that have fencing disabled. Even disabling it for a test cluster means you won't be able to test real failure scenarios. .. index:: single: fencing; device Choose a Fence Device ##################### The two broad categories of fence device are power fencing, which cuts off power to the target, and fabric fencing, which cuts off the target's access to some critical resource, such as a shared disk or access to the local network. Power fencing devices include: * Intelligent power switches * IPMI * Hardware watchdog device (alone, or in combination with shared storage used as a "poison pill" mechanism) Fabric fencing devices include: * Shared storage that can be cut off for a target host by another host (for example, an external storage device that supports SCSI-3 persistent reservations) * Intelligent network switches Using IPMI as a power fencing device may seem like a good choice. However, if the IPMI shares power and/or network access with the host (such as most onboard IPMI controllers), a power or network failure will cause both the host and its fencing device to fail. The cluster will be unable to recover, and must stop all resources to avoid a possible split-brain situation. Likewise, any device that relies on the machine being active (such as SSH-based "devices" sometimes used during testing) is inappropriate, because fencing will be required when the node is completely unresponsive. (Fence agents like ``fence_ilo_ssh``, which connects via SSH to an HP iLO but not to the cluster node, are fine.) Configure the Cluster for Fencing ################################# #. Install the fence agent(s). To see what packages are available, run ``dnf search fence-``. Be sure to install the package(s) on all cluster nodes. #. Configure the fence device itself to be able to fence your nodes and accept fencing requests. This includes any necessary configuration on the device and on the nodes, and any firewall or SELinux changes needed. Test the communication between the device and your nodes. #. Find the name of the correct fence agent: ``pcs stonith list`` #. Find the parameters associated with the device: ``pcs stonith describe `` #. Create a local copy of the CIB: ``pcs cluster cib stonith_cfg`` #. Create the fencing resource: ``pcs -f stonith_cfg stonith create [STONITH_DEVICE_OPTIONS]`` Any flags that do not take arguments, such as ``--ssl``, should be passed as ``ssl=1``. #. Ensure fencing is enabled in the cluster: ``pcs -f stonith_cfg property set stonith-enabled=true`` #. If the device does not know how to fence nodes based on their cluster node name, you may also need to set the special ``pcmk_host_map`` parameter. See ``man pacemaker-fenced`` for details. #. If the device does not support the ``list`` command, you may also need to set the special ``pcmk_host_list`` and/or ``pcmk_host_check`` parameters. See ``man pacemaker-fenced`` for details. -#. If the device does not expect the victim to be specified with the ``port`` +#. If the device does not expect the target to be specified with the ``port`` parameter, you may also need to set the special ``pcmk_host_argument`` parameter. See ``man pacemaker-fenced`` for details. #. Commit the new configuration: ``pcs cluster cib-push stonith_cfg`` #. Once the fence device resource is running, test it (you might want to stop the cluster on that machine first): ``pcs stonith fence `` Example ####### For this example, assume we have a chassis containing four nodes and a separately powered IPMI device active on ``10.0.0.1``. Following the steps above would go something like this: Step 1: Install the ``fence-agents-ipmilan`` package on both nodes. Step 2: Configure the IP address, authentication credentials, etc. in the IPMI device itself. Step 3: Choose the ``fence_ipmilan`` STONITH agent. Step 4: Obtain the agent's possible parameters: .. code-block:: console [root@pcmk-1 ~]# pcs stonith describe fence_ipmilan fence_ipmilan - Fence agent for IPMI fence_ipmilan is an I/O Fencing agentwhich can be used with machines controlled by IPMI.This agent calls support software ipmitool (http://ipmitool.sf.net/). WARNING! This fence agent might report success before the node is powered off. You should use -m/method onoff if your fence device works correctly with that option. Stonith options: auth: IPMI Lan Auth type. cipher: Ciphersuite to use (same as ipmitool -C parameter) hexadecimal_kg: Hexadecimal-encoded Kg key for IPMIv2 authentication ip: IP address or hostname of fencing device ipport: TCP/UDP port to use for connection with device lanplus: Use Lanplus to improve security of connection method: Method to fence password: Login password or passphrase password_script: Script to run to retrieve password plug: IP address or hostname of fencing device (together with --port-as-ip) privlvl: Privilege level on IPMI device target: Bridge IPMI requests to the remote target address username: Login name quiet: Disable logging to stderr. Does not affect --verbose or --debug-file or logging to syslog. verbose: Verbose mode. Multiple -v flags can be stacked on the command line (e.g., -vvv) to increase verbosity. verbose_level: Level of debugging detail in output. Defaults to the number of --verbose flags specified on the command line, or to 1 if verbose=1 in a stonith device configuration (i.e., on stdin). debug_file: Write debug information to given file delay: Wait X seconds before fencing is started disable_timeout: Disable timeout (true/false) (default: true when run from Pacemaker 2.0+) ipmitool_path: Path to ipmitool binary login_timeout: Wait X seconds for cmd prompt after login port_as_ip: Make "port/plug" to be an alias to IP address power_timeout: Test X seconds for status change after ON/OFF power_wait: Wait X seconds after issuing ON/OFF shell_timeout: Wait X seconds for cmd prompt after issuing command stonith_status_sleep: Sleep X seconds between status calls during a STONITH action ipmitool_timeout: Timeout (sec) for IPMI operation retry_on: Count of attempts to retry power on use_sudo: Use sudo (without password) when calling 3rd party software sudo_path: Path to sudo binary pcmk_host_map: A mapping of host names to ports numbers for devices that do not support host names. Eg. node1:1;node2:2,3 would tell the cluster to use port 1 for node1 and ports 2 and 3 for node2 pcmk_host_list: A list of machines controlled by this device (Optional unless pcmk_host_check=static-list). pcmk_host_check: How to determine which machines are controlled by the device. Allowed values: dynamic-list (query the device via the 'list' command), static-list (check the pcmk_host_list attribute), status (query the device via the 'status' command), none (assume every device can fence every machine) pcmk_delay_max: Enable a delay of no more than the time specified before executing fencing actions. Pacemaker derives the overall delay by taking the value of pcmk_delay_base and adding a random delay value such that the sum is kept below this maximum. This prevents double fencing when using slow devices such as sbd. Use this to enable a random delay for fencing actions. The overall delay is derived from this random delay value adding a static delay so that the sum is kept below the maximum delay. pcmk_delay_base: Enable a base delay for fencing actions and specify base delay value. This enables a static delay for fencing actions, which can help avoid "death matches" where two nodes try to fence each other at the same time. If pcmk_delay_max is also used, a random delay will be added such that the total delay is kept below that value. This can be set to a single time value to apply to any node targeted by this device (useful if a separate device is configured for each target), or to a node map (for example, "node1:1s;node2:5") to set a different value per target. pcmk_action_limit: The maximum number of actions can be performed in parallel on this device Cluster property concurrent-fencing=true needs to be configured first. Then use this to specify the maximum number of actions can be performed in parallel on this device. -1 is unlimited. Default operations: monitor: interval=60s Step 5: ``pcs cluster cib stonith_cfg`` Step 6: Here are example parameters for creating our fence device resource: .. code-block:: console [root@pcmk-1 ~]# pcs -f stonith_cfg stonith create ipmi-fencing fence_ipmilan \ pcmk_host_list="pcmk-1 pcmk-2" ipaddr=10.0.0.1 login=testuser \ passwd=acd123 op monitor interval=60s [root@pcmk-1 ~]# pcs -f stonith_cfg stonith * ipmi-fencing (stonith:fence_ipmilan): Stopped Steps 7-10: Enable fencing in the cluster: .. code-block:: console [root@pcmk-1 ~]# pcs -f stonith_cfg property set stonith-enabled=true [root@pcmk-1 ~]# pcs -f stonith_cfg property Cluster Properties: cluster-infrastructure: corosync cluster-name: mycluster dc-version: 2.0.5-4.el8-ba59be7122 have-watchdog: false stonith-enabled: true Step 11: ``pcs cluster cib-push stonith_cfg --config`` Step 12: Test: .. code-block:: console [root@pcmk-1 ~]# pcs cluster stop pcmk-2 [root@pcmk-1 ~]# pcs stonith fence pcmk-2 After a successful test, login to any rebooted nodes, and start the cluster (with ``pcs cluster start``). .. [#] If the data is corrupt, there is little point in continuing to make it available. diff --git a/doc/sphinx/Pacemaker_Explained/fencing.rst b/doc/sphinx/Pacemaker_Explained/fencing.rst index 4e8d3a2f78..109b4da604 100644 --- a/doc/sphinx/Pacemaker_Explained/fencing.rst +++ b/doc/sphinx/Pacemaker_Explained/fencing.rst @@ -1,1298 +1,1298 @@ .. index:: single: fencing single: STONITH .. _fencing: Fencing ------- What Is Fencing? ################ *Fencing* is the ability to make a node unable to run resources, even when that node is unresponsive to cluster commands. Fencing is also known as *STONITH*, an acronym for "Shoot The Other Node In The Head", since the most common fencing method is cutting power to the node. Another method is "fabric fencing", cutting the node's access to some capability required to run resources (such as network access or a shared disk). .. index:: single: fencing; why necessary Why Is Fencing Necessary? ######################### Fencing protects your data from being corrupted by malfunctioning nodes or unintentional concurrent access to shared resources. Fencing protects against the "split brain" failure scenario, where cluster nodes have lost the ability to reliably communicate with each other but are still able to run resources. If the cluster just assumed that uncommunicative nodes were down, then multiple instances of a resource could be started on different nodes. The effect of split brain depends on the resource type. For example, an IP address brought up on two hosts on a network will cause packets to randomly be sent to one or the other host, rendering the IP useless. For a database or clustered file system, the effect could be much more severe, causing data corruption or divergence. Fencing is also used when a resource cannot otherwise be stopped. If a resource fails to stop on a node, it cannot be started on a different node without risking the same type of conflict as split-brain. Fencing the original node ensures the resource can be safely started elsewhere. Users may also configure the ``on-fail`` property of :ref:`operation` or the ``loss-policy`` property of :ref:`ticket constraints ` to ``fence``, in which case the cluster will fence the resource's node if the operation fails or the ticket is lost. .. index:: single: fencing; device Fence Devices ############# A *fence device* or *fencing device* is a special type of resource that provides the means to fence a node. Examples of fencing devices include intelligent power switches and IPMI devices that accept SNMP commands to cut power to a node, and iSCSI controllers that allow SCSI reservations to be used to cut a node's access to a shared disk. Since fencing devices will be used to recover from loss of networking connectivity to other nodes, it is essential that they do not rely on the same network as the cluster itself, otherwise that network becomes a single point of failure. Since loss of a node due to power outage is indistinguishable from loss of network connectivity to that node, it is also essential that at least one fence device for a node does not share power with that node. For example, an on-board IPMI controller that shares power with its host should not be used as the sole fencing device for that host. Since fencing is used to isolate malfunctioning nodes, no fence device should rely on its target functioning properly. This includes, for example, devices that ssh into a node and issue a shutdown command (such devices might be suitable for testing, but never for production). .. index:: single: fencing; agent Fence Agents ############ A *fence agent* or *fencing agent* is a ``stonith``-class resource agent. The fence agent standard provides commands (such as ``off`` and ``reboot``) that the cluster can use to fence nodes. As with other resource agent classes, this allows a layer of abstraction so that Pacemaker doesn't need any knowledge about specific fencing technologies -- that knowledge is isolated in the agent. Pacemaker supports two fence agent standards, both inherited from no-longer-active projects: * Red Hat Cluster Suite (RHCS) style: These are typically installed in ``/usr/sbin`` with names starting with ``fence_``. * Linux-HA style: These typically have names starting with ``external/``. Pacemaker can support these agents using the **fence_legacy** RHCS-style agent as a wrapper, *if* support was enabled when Pacemaker was built, which requires the ``cluster-glue`` library. When a Fence Device Can Be Used ############################### Fencing devices do not actually "run" like most services. Typically, they just provide an interface for sending commands to an external device. Additionally, fencing may be initiated by Pacemaker, by other cluster-aware software such as DRBD or DLM, or manually by an administrator, at any point in the cluster life cycle, including before any resources have been started. To accommodate this, Pacemaker does not require the fence device resource to be "started" in order to be used. Whether a fence device is started or not determines whether a node runs any recurring monitor for the device, and gives the node a slight preference for being chosen to execute fencing using that device. By default, any node can execute any fencing device. If a fence device is disabled by setting its ``target-role`` to ``Stopped``, then no node can use that device. If a location constraint with a negative score prevents a specific node from "running" a fence device, then that node will never be chosen to execute fencing using the device. A node may fence itself, but the cluster will choose that only if no other nodes can do the fencing. A common configuration scenario is to have one fence device per target node. In such a case, users often configure anti-location constraints so that the target node does not monitor its own device. Limitations of Fencing Resources ################################ Fencing resources have certain limitations that other resource classes don't: * They may have only one set of meta-attributes and one set of instance attributes. * If :ref:`rules` are used to determine fencing resource options, these might be evaluated only when first read, meaning that later changes to the rules will have no effect. Therefore, it is better to avoid confusion and not use rules at all with fencing resources. These limitations could be revisited if there is sufficient user demand. .. index:: single: fencing; special instance attributes .. _fencing-attributes: Special Meta-Attributes for Fencing Resources ############################################# The table below lists special resource meta-attributes that may be set for any fencing resource. .. table:: **Additional Properties of Fencing Resources** :widths: 2 1 2 4 +----------------------+---------+--------------------+----------------------------------------+ | Field | Type | Default | Description | +======================+=========+====================+========================================+ | provides | string | | .. index:: | | | | | single: provides | | | | | | | | | | Any special capability provided by the | | | | | fence device. Currently, only one such | | | | | capability is meaningful: | | | | | :ref:`unfencing `. | +----------------------+---------+--------------------+----------------------------------------+ Special Instance Attributes for Fencing Resources ################################################# The table below lists special instance attributes that may be set for any fencing resource (*not* meta-attributes, even though they are interpreted by Pacemaker rather than the fence agent). These are also listed in the man page for ``pacemaker-fenced``. .. Not_Yet_Implemented: +----------------------+---------+--------------------+----------------------------------------+ | priority | integer | 0 | .. index:: | | | | | single: priority | | | | | | | | | | The priority of the fence device. | | | | | Devices are tried in order of highest | | | | | priority to lowest. | +----------------------+---------+--------------------+----------------------------------------+ .. table:: **Additional Properties of Fencing Resources** :class: longtable :widths: 2 1 2 4 +----------------------+---------+--------------------+----------------------------------------+ | Field | Type | Default | Description | +======================+=========+====================+========================================+ | stonith-timeout | time | | .. index:: | | | | | single: stonith-timeout | | | | | | | | | | This is not used by Pacemaker (see the | | | | | ``pcmk_reboot_timeout``, | | | | | ``pcmk_off_timeout``, etc. properties | | | | | instead), but it may be used by | | | | | Linux-HA fence agents. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_host_map | string | | .. index:: | | | | | single: pcmk_host_map | | | | | | | | | | A mapping of node names to ports | | | | | for devices that do not understand | | | | | the node names. | | | | | | | | | | Example: ``node1:1;node2:2,3`` tells | | | | | the cluster to use port 1 for | | | | | ``node1`` and ports 2 and 3 for | | | | | ``node2``. If ``pcmk_host_check`` is | | | | | explicitly set to ``static-list``, | | | | | either this or ``pcmk_host_list`` must | | | | | be set. The port portion of the map | | | | | may contain special characters such as | | | | | spaces if preceded by a backslash | | | | | *(since 2.1.2)*. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_host_list | string | | .. index:: | | | | | single: pcmk_host_list | | | | | | | | | | A list of machines controlled by this | | | | | device. If ``pcmk_host_check`` is | | | | | explicitly set to ``static-list``, | | | | | either this or ``pcmk_host_map`` must | | | | | be set. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_host_check | string | Value appropriate | .. index:: | | | | to other | single: pcmk_host_check | | | | parameters (see | | | | | "Default Check | The method Pacemaker should use to | | | | Type" below) | determine which nodes can be targeted | | | | | by this device. Allowed values: | | | | | | | | | | * ``static-list:`` targets are listed | | | | | in the ``pcmk_host_list`` or | | | | | ``pcmk_host_map`` attribute | | | | | * ``dynamic-list:`` query the device | | | | | via the agent's ``list`` action | | | | | * ``status:`` query the device via the | | | | | agent's ``status`` action | | | | | * ``none:`` assume the device can | | | | | fence any node | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_delay_max | time | 0s | .. index:: | | | | | single: pcmk_delay_max | | | | | | | | | | Enable a delay of no more than the | | | | | time specified before executing | | | | | fencing actions. Pacemaker derives the | | | | | overall delay by taking the value of | | | | | pcmk_delay_base and adding a random | | | | | delay value such that the sum is kept | | | | | below this maximum. This is sometimes | | | | | used in two-node clusters to ensure | | | | | that the nodes don't fence each other | | | | | at the same time. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_delay_base | time | 0s | .. index:: | | | | | single: pcmk_delay_base | | | | | | | | | | Enable a static delay before executing | | | | | fencing actions. This can be used, for | | | | | example, in two-node clusters to | | | | | ensure that the nodes don't fence each | | | | | other, by having separate fencing | | | | | resources with different values. The | | | | | node that is fenced with the shorter | | | | | delay will lose a fencing race. The | | | | | overall delay introduced by pacemaker | | | | | is derived from this value plus a | | | | | random delay such that the sum is kept | | | | | below the maximum delay. A single | | | | | device can have different delays per | | | | | node using a host map *(since 2.1.2)*, | | | | | for example ``node1:0s;node2:5s.`` | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_action_limit | integer | 1 | .. index:: | | | | | single: pcmk_action_limit | | | | | | | | | | The maximum number of actions that can | | | | | be performed in parallel on this | | | | | device. A value of -1 means unlimited. | | | | | Node fencing actions initiated by the | | | | | cluster (as opposed to an administrator| | | | | running the ``stonith_admin`` tool or | | | | | the fencer running recurring device | | | | | monitors and ``status`` and ``list`` | | | | | commands) are additionally subject to | | | | | the ``concurrent-fencing`` cluster | | | | | property. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_host_argument | string | ``port`` otherwise | .. index:: | | | | ``plug`` if | single: pcmk_host_argument | | | | supported | | | | | according to the | *Advanced use only.* Which parameter | | | | metadata of the | should be supplied to the fence agent | | | | fence agent | to identify the node to be fenced. | | | | | Some devices support neither the | | | | | standard ``plug`` nor the deprecated | | | | | ``port`` parameter, or may provide | | | | | additional ones. Use this to specify | | | | | an alternate, device-specific | | | | | parameter. A value of ``none`` tells | | | | | the cluster not to supply any | | | | | additional parameters. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_reboot_action | string | reboot | .. index:: | | | | | single: pcmk_reboot_action | | | | | | | | | | *Advanced use only.* The command to | | | | | send to the resource agent in order to | | | | | reboot a node. Some devices do not | | | | | support the standard commands or may | | | | | provide additional ones. Use this to | | | | | specify an alternate, device-specific | | | | | command. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_reboot_timeout | time | 60s | .. index:: | | | | | single: pcmk_reboot_timeout | | | | | | | | | | *Advanced use only.* Specify an | | | | | alternate timeout to use for | | | | | ``reboot`` actions instead of the | | | | | value of ``stonith-timeout``. Some | | | | | devices need much more or less time to | | | | | complete than normal. Use this to | | | | | specify an alternate, device-specific | | | | | timeout. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_reboot_retries | integer | 2 | .. index:: | | | | | single: pcmk_reboot_retries | | | | | | | | | | *Advanced use only.* The maximum | | | | | number of times to retry the | | | | | ``reboot`` command within the timeout | | | | | period. Some devices do not support | | | | | multiple connections, and operations | | | | | may fail if the device is busy with | | | | | another task, so Pacemaker will | | | | | automatically retry the operation, if | | | | | there is time remaining. Use this | | | | | option to alter the number of times | | | | | Pacemaker retries before giving up. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_off_action | string | off | .. index:: | | | | | single: pcmk_off_action | | | | | | | | | | *Advanced use only.* The command to | | | | | send to the resource agent in order to | | | | | shut down a node. Some devices do not | | | | | support the standard commands or may | | | | | provide additional ones. Use this to | | | | | specify an alternate, device-specific | | | | | command. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_off_timeout | time | 60s | .. index:: | | | | | single: pcmk_off_timeout | | | | | | | | | | *Advanced use only.* Specify an | | | | | alternate timeout to use for | | | | | ``off`` actions instead of the | | | | | value of ``stonith-timeout``. Some | | | | | devices need much more or less time to | | | | | complete than normal. Use this to | | | | | specify an alternate, device-specific | | | | | timeout. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_off_retries | integer | 2 | .. index:: | | | | | single: pcmk_off_retries | | | | | | | | | | *Advanced use only.* The maximum | | | | | number of times to retry the | | | | | ``off`` command within the timeout | | | | | period. Some devices do not support | | | | | multiple connections, and operations | | | | | may fail if the device is busy with | | | | | another task, so Pacemaker will | | | | | automatically retry the operation, if | | | | | there is time remaining. Use this | | | | | option to alter the number of times | | | | | Pacemaker retries before giving up. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_list_action | string | list | .. index:: | | | | | single: pcmk_list_action | | | | | | | | | | *Advanced use only.* The command to | | | | | send to the resource agent in order to | | | | | list nodes. Some devices do not | | | | | support the standard commands or may | | | | | provide additional ones. Use this to | | | | | specify an alternate, device-specific | | | | | command. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_list_timeout | time | 60s | .. index:: | | | | | single: pcmk_list_timeout | | | | | | | | | | *Advanced use only.* Specify an | | | | | alternate timeout to use for | | | | | ``list`` actions instead of the | | | | | value of ``stonith-timeout``. Some | | | | | devices need much more or less time to | | | | | complete than normal. Use this to | | | | | specify an alternate, device-specific | | | | | timeout. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_list_retries | integer | 2 | .. index:: | | | | | single: pcmk_list_retries | | | | | | | | | | *Advanced use only.* The maximum | | | | | number of times to retry the | | | | | ``list`` command within the timeout | | | | | period. Some devices do not support | | | | | multiple connections, and operations | | | | | may fail if the device is busy with | | | | | another task, so Pacemaker will | | | | | automatically retry the operation, if | | | | | there is time remaining. Use this | | | | | option to alter the number of times | | | | | Pacemaker retries before giving up. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_monitor_action | string | monitor | .. index:: | | | | | single: pcmk_monitor_action | | | | | | | | | | *Advanced use only.* The command to | | | | | send to the resource agent in order to | | | | | report extended status. Some devices do| | | | | not support the standard commands or | | | | | may provide additional ones. Use this | | | | | to specify an alternate, | | | | | device-specific command. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_monitor_timeout | time | 60s | .. index:: | | | | | single: pcmk_monitor_timeout | | | | | | | | | | *Advanced use only.* Specify an | | | | | alternate timeout to use for | | | | | ``monitor`` actions instead of the | | | | | value of ``stonith-timeout``. Some | | | | | devices need much more or less time to | | | | | complete than normal. Use this to | | | | | specify an alternate, device-specific | | | | | timeout. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_monitor_retries | integer | 2 | .. index:: | | | | | single: pcmk_monitor_retries | | | | | | | | | | *Advanced use only.* The maximum | | | | | number of times to retry the | | | | | ``monitor`` command within the timeout | | | | | period. Some devices do not support | | | | | multiple connections, and operations | | | | | may fail if the device is busy with | | | | | another task, so Pacemaker will | | | | | automatically retry the operation, if | | | | | there is time remaining. Use this | | | | | option to alter the number of times | | | | | Pacemaker retries before giving up. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_status_action | string | status | .. index:: | | | | | single: pcmk_status_action | | | | | | | | | | *Advanced use only.* The command to | | | | | send to the resource agent in order to | | | | | report status. Some devices do | | | | | not support the standard commands or | | | | | may provide additional ones. Use this | | | | | to specify an alternate, | | | | | device-specific command. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_status_timeout | time | 60s | .. index:: | | | | | single: pcmk_status_timeout | | | | | | | | | | *Advanced use only.* Specify an | | | | | alternate timeout to use for | | | | | ``status`` actions instead of the | | | | | value of ``stonith-timeout``. Some | | | | | devices need much more or less time to | | | | | complete than normal. Use this to | | | | | specify an alternate, device-specific | | | | | timeout. | +----------------------+---------+--------------------+----------------------------------------+ | pcmk_status_retries | integer | 2 | .. index:: | | | | | single: pcmk_status_retries | | | | | | | | | | *Advanced use only.* The maximum | | | | | number of times to retry the | | | | | ``status`` command within the timeout | | | | | period. Some devices do not support | | | | | multiple connections, and operations | | | | | may fail if the device is busy with | | | | | another task, so Pacemaker will | | | | | automatically retry the operation, if | | | | | there is time remaining. Use this | | | | | option to alter the number of times | | | | | Pacemaker retries before giving up. | +----------------------+---------+--------------------+----------------------------------------+ Default Check Type ################## If the user does not explicitly configure ``pcmk_host_check`` for a fence device, a default value appropriate to other configured parameters will be used: * If either ``pcmk_host_list`` or ``pcmk_host_map`` is configured, ``static-list`` will be used; * otherwise, if the fence device supports the ``list`` action, and the first attempt at using ``list`` succeeds, ``dynamic-list`` will be used; * otherwise, if the fence device supports the ``status`` action, ``status`` will be used; * otherwise, ``none`` will be used. .. index:: single: unfencing single: fencing; unfencing .. _unfencing: Unfencing ######### With fabric fencing (such as cutting network or shared disk access rather than power), it is expected that the cluster will fence the node, and then a system administrator must manually investigate what went wrong, correct any issues found, then reboot (or restart the cluster services on) the node. Once the node reboots and rejoins the cluster, some fabric fencing devices require an explicit command to restore the node's access. This capability is called *unfencing* and is typically implemented as the fence agent's ``on`` command. If any cluster resource has ``requires`` set to ``unfencing``, then that resource will not be probed or started on a node until that node has been unfenced. Fencing and Quorum ################## In general, a cluster partition may execute fencing only if the partition has quorum, and the ``stonith-enabled`` cluster property is set to true. However, there are exceptions: * The requirements apply only to fencing initiated by Pacemaker. If an administrator initiates fencing using the ``stonith_admin`` command, or an external application such as DLM initiates fencing using Pacemaker's C API, the requirements do not apply. * A cluster partition without quorum is allowed to fence any active member of that partition. As a corollary, this allows a ``no-quorum-policy`` of ``suicide`` to work. * If the ``no-quorum-policy`` cluster property is set to ``ignore``, then quorum is not required to execute fencing of any node. Fencing Timeouts ################ Fencing timeouts are complicated, since a single fencing operation can involve many steps, each of which may have a separate timeout. Fencing may be initiated in one of several ways: * An administrator may initiate fencing using the ``stonith_admin`` tool, which has a ``--timeout`` option (defaulting to 2 minutes) that will be used as the fence operation timeout. * An external application such as DLM may initiate fencing using the Pacemaker C API. The application will specify the fence operation timeout in this case, which might or might not be configurable by the user. * The cluster may initiate fencing itself. In this case, the ``stonith-timeout`` cluster property (defaulting to 1 minute) will be used as the fence operation timeout. However fencing is initiated, the initiator contacts Pacemaker's fencer (``pacemaker-fenced``) to request fencing. This connection and request has its own timeout, separate from the fencing operation timeout, but usually happens very quickly. The fencer will contact all fencers in the cluster to ask what devices they have available to fence the target node. The fence operation timeout will be used as the timeout for each of these queries. Once a fencing device has been selected, the fencer will check whether any action-specific timeout has been configured for the device, to use instead of the fence operation timeout. For example, if ``stonith-timeout`` is 60 seconds, but the fencing device has ``pcmk_reboot_timeout`` configured as 90 seconds, then a timeout of 90 seconds will be used for reboot actions using that device. A device may have retries configured, in which case the timeout applies across all attempts. For example, if a device has ``pcmk_reboot_retries`` configured as 2, and the first reboot attempt fails, the second attempt will only have whatever time is remaining in the action timeout after subtracting how much time the first attempt used. This means that if the first attempt fails due to using the entire timeout, no further attempts will be made. There is currently no way to configure a per-attempt timeout. If more than one device is required to fence a target, whether due to failure of the first device or a fencing topology with multiple devices configured for the target, each device will have its own separate action timeout. For all of the above timeouts, the fencer will generally multiply the configured value by 1.2 to get an actual value to use, to account for time needed by the fencer's own processing. Separate from the fencer's timeouts, some fence agents have internal timeouts for individual steps of their fencing process. These agents often have parameters to configure these timeouts, such as ``login-timeout``, ``shell-timeout``, or ``power-timeout``. Many such agents also have a ``disable-timeout`` parameter to ignore their internal timeouts and just let Pacemaker handle the timeout. This causes a difference in retry behavior. If ``disable-timeout`` is not set, and the agent hits one of its internal timeouts, it will report that as a failure to Pacemaker, which can then retry. If ``disable-timeout`` is set, and Pacemaker hits a timeout for the agent, then there will be no time remaining, and no retry will be done. Fence Devices Dependent on Other Resources ########################################## In some cases, a fence device may require some other cluster resource (such as an IP address) to be active in order to function properly. This is obviously undesirable in general: fencing may be required when the depended-on resource is not active, or fencing may be required because the node running the depended-on resource is no longer responding. However, this may be acceptable under certain conditions: * The dependent fence device should not be able to target any node that is allowed to run the depended-on resource. * The depended-on resource should not be disabled during production operation. * The ``concurrent-fencing`` cluster property should be set to ``true``. Otherwise, if both the node running the depended-on resource and some node targeted by the dependent fence device need to be fenced, the fencing of the node running the depended-on resource might be ordered first, making the second fencing impossible and blocking further recovery. With concurrent fencing, the dependent fence device might fail at first due to the depended-on resource being unavailable, but it will be retried and eventually succeed once the resource is brought back up. Even under those conditions, there is one unlikely problem scenario. The DC always schedules fencing of itself after any other fencing needed, to avoid unnecessary repeated DC elections. If the dependent fence device targets the DC, and both the DC and a different node running the depended-on resource need to be fenced, the DC fencing will always fail and block further recovery. Note, however, that losing a DC node entirely causes some other node to become DC and schedule the fencing, so this is only a risk when a stop or other operation with ``on-fail`` set to ``fencing`` fails on the DC. .. index:: single: fencing; configuration Configuring Fencing ################### Higher-level tools can provide simpler interfaces to this process, but using Pacemaker command-line tools, this is how you could configure a fence device. #. Find the correct driver: .. code-block:: none # stonith_admin --list-installed .. note:: You may have to install packages to make fence agents available on your host. Searching your available packages for ``fence-`` is usually helpful. Ensure the packages providing the fence agents you require are installed on every cluster node. #. Find the required parameters associated with the device (replacing ``$AGENT_NAME`` with the name obtained from the previous step): .. code-block:: none # stonith_admin --metadata --agent $AGENT_NAME #. Create a file called ``stonith.xml`` containing a primitive resource with a class of ``stonith``, a type equal to the agent name obtained earlier, and a parameter for each of the values returned in the previous step. #. If the device does not know how to fence nodes based on their uname, you may also need to set the special ``pcmk_host_map`` parameter. See :ref:`fencing-attributes` for details. #. If the device does not support the ``list`` command, you may also need to set the special ``pcmk_host_list`` and/or ``pcmk_host_check`` parameters. See :ref:`fencing-attributes` for details. -#. If the device does not expect the victim to be specified with the +#. If the device does not expect the target to be specified with the ``port`` parameter, you may also need to set the special ``pcmk_host_argument`` parameter. See :ref:`fencing-attributes` for details. #. Upload it into the CIB using cibadmin: .. code-block:: none # cibadmin --create --scope resources --xml-file stonith.xml #. Set ``stonith-enabled`` to true: .. code-block:: none # crm_attribute --type crm_config --name stonith-enabled --update true #. Once the stonith resource is running, you can test it by executing the following, replacing ``$NODE_NAME`` with the name of the node to fence (although you might want to stop the cluster on that machine first): .. code-block:: none # stonith_admin --reboot $NODE_NAME Example Fencing Configuration _____________________________ For this example, we assume we have a cluster node, ``pcmk-1``, whose IPMI controller is reachable at the IP address 192.0.2.1. The IPMI controller uses the username ``testuser`` and the password ``abc123``. #. Looking at what's installed, we may see a variety of available agents: .. code-block:: none # stonith_admin --list-installed .. code-block:: none (... some output omitted ...) fence_idrac fence_ilo3 fence_ilo4 fence_ilo5 fence_imm fence_ipmilan (... some output omitted ...) Perhaps after some reading some man pages and doing some Internet searches, we might decide ``fence_ipmilan`` is our best choice. #. Next, we would check what parameters ``fence_ipmilan`` provides: .. code-block:: none # stonith_admin --metadata -a fence_ipmilan .. code-block:: xml fence_ipmilan is an I/O Fencing agentwhich can be used with machines controlled by IPMI.This agent calls support software ipmitool (http://ipmitool.sf.net/). WARNING! This fence agent might report success before the node is powered off. You should use -m/method onoff if your fence device works correctly with that option. Fencing action IPMI Lan Auth type. Ciphersuite to use (same as ipmitool -C parameter) Hexadecimal-encoded Kg key for IPMIv2 authentication IP address or hostname of fencing device IP address or hostname of fencing device TCP/UDP port to use for connection with device Use Lanplus to improve security of connection Login name Method to fence Login password or passphrase Script to run to retrieve password Login password or passphrase Script to run to retrieve password IP address or hostname of fencing device (together with --port-as-ip) IP address or hostname of fencing device (together with --port-as-ip) Privilege level on IPMI device Bridge IPMI requests to the remote target address Login name Disable logging to stderr. Does not affect --verbose or --debug-file or logging to syslog. Verbose mode Write debug information to given file Write debug information to given file Display version information and exit Display help and exit Wait X seconds before fencing is started Path to ipmitool binary Wait X seconds for cmd prompt after login Make "port/plug" to be an alias to IP address Test X seconds for status change after ON/OFF Wait X seconds after issuing ON/OFF Wait X seconds for cmd prompt after issuing command Count of attempts to retry power on Use sudo (without password) when calling 3rd party software Use sudo (without password) when calling 3rd party software Path to sudo binary Once we've decided what parameter values we think we need, it is a good idea to run the fence agent's status action manually, to verify that our values work correctly: .. code-block:: none # fence_ipmilan --lanplus -a 192.0.2.1 -l testuser -p abc123 -o status Chassis Power is on #. Based on that, we might create a fencing resource configuration like this in ``stonith.xml`` (or any file name, just use the same name with ``cibadmin`` later): .. code-block:: xml .. note:: Even though the man page shows that the ``action`` parameter is supported, we do not provide that in the resource configuration. Pacemaker will supply an appropriate action whenever the fence device must be used. #. In this case, we don't need to configure ``pcmk_host_map`` because ``fence_ipmilan`` ignores the target node name and instead uses its ``ip`` parameter to know how to contact the IPMI controller. #. We do need to let Pacemaker know which cluster node can be fenced by this device, since ``fence_ipmilan`` doesn't support the ``list`` action. Add a line like this to the agent's instance attributes: .. code-block:: xml #. We don't need to configure ``pcmk_host_argument`` since ``ip`` is all the fence agent needs (it ignores the target name). #. Make the configuration active: .. code-block:: none # cibadmin --create --scope resources --xml-file stonith.xml #. Set ``stonith-enabled`` to true (this only has to be done once): .. code-block:: none # crm_attribute --type crm_config --name stonith-enabled --update true #. Since our cluster is still in testing, we can reboot ``pcmk-1`` without bothering anyone, so we'll test our fencing configuration by running this from one of the other cluster nodes: .. code-block:: none # stonith_admin --reboot pcmk-1 Then we will verify that the node did, in fact, reboot. We can repeat that process to create a separate fencing resource for each node. With some other fence device types, a single fencing resource is able to be used for all nodes. In fact, we could do that with ``fence_ipmilan``, using the ``port-as-ip`` parameter along with ``pcmk_host_map``. Either approach is fine. .. index:: single: fencing; topology single: fencing-topology single: fencing-level Fencing Topologies ################## Pacemaker supports fencing nodes with multiple devices through a feature called *fencing topologies*. Fencing topologies may be used to provide alternative devices in case one fails, or to require multiple devices to all be executed successfully in order to consider the node successfully fenced, or even a combination of the two. Create the individual devices as you normally would, then define one or more ``fencing-level`` entries in the ``fencing-topology`` section of the configuration. * Each fencing level is attempted in order of ascending ``index``. Allowed values are 1 through 9. * If a device fails, processing terminates for the current level. No further devices in that level are exercised, and the next level is attempted instead. * If the operation succeeds for all the listed devices in a level, the level is deemed to have passed. * The operation is finished when a level has passed (success), or all levels have been attempted (failed). * If the operation failed, the next step is determined by the scheduler and/or the controller. Some possible uses of topologies include: * Try on-board IPMI, then an intelligent power switch if that fails * Try fabric fencing of both disk and network, then fall back to power fencing if either fails * Wait up to a certain time for a kernel dump to complete, then cut power to the node .. table:: **Attributes of a fencing-level Element** :class: longtable :widths: 1 4 +------------------+-----------------------------------------------------------------------------------------+ | Attribute | Description | +==================+=========================================================================================+ | id | .. index:: | | | pair: fencing-level; id | | | | | | A unique name for this element (required) | +------------------+-----------------------------------------------------------------------------------------+ | target | .. index:: | | | pair: fencing-level; target | | | | | | The name of a single node to which this level applies | +------------------+-----------------------------------------------------------------------------------------+ | target-pattern | .. index:: | | | pair: fencing-level; target-pattern | | | | | | An extended regular expression (as defined in `POSIX | | | `_) | | | matching the names of nodes to which this level applies | +------------------+-----------------------------------------------------------------------------------------+ | target-attribute | .. index:: | | | pair: fencing-level; target-attribute | | | | | | The name of a node attribute that is set (to ``target-value``) for nodes to which this | | | level applies | +------------------+-----------------------------------------------------------------------------------------+ | target-value | .. index:: | | | pair: fencing-level; target-value | | | | | | The node attribute value (of ``target-attribute``) that is set for nodes to which this | | | level applies | +------------------+-----------------------------------------------------------------------------------------+ | index | .. index:: | | | pair: fencing-level; index | | | | | | The order in which to attempt the levels. Levels are attempted in ascending order | | | *until one succeeds*. Valid values are 1 through 9. | +------------------+-----------------------------------------------------------------------------------------+ | devices | .. index:: | | | pair: fencing-level; devices | | | | | | A comma-separated list of devices that must all be tried for this level | +------------------+-----------------------------------------------------------------------------------------+ .. note:: **Fencing topology with different devices for different nodes** .. code-block:: xml ... ... Example Dual-Layer, Dual-Device Fencing Topologies __________________________________________________ The following example illustrates an advanced use of ``fencing-topology`` in a cluster with the following properties: * 2 nodes (prod-mysql1 and prod-mysql2) * the nodes have IPMI controllers reachable at 192.0.2.1 and 192.0.2.2 * the nodes each have two independent Power Supply Units (PSUs) connected to two independent Power Distribution Units (PDUs) reachable at 198.51.100.1 (port 10 and port 11) and 203.0.113.1 (port 10 and port 11) * fencing via the IPMI controller uses the ``fence_ipmilan`` agent (1 fence device per controller, with each device targeting a separate node) * fencing via the PDUs uses the ``fence_apc_snmp`` agent (1 fence device per PDU, with both devices targeting both nodes) * a random delay is used to lessen the chance of a "death match" * fencing topology is set to try IPMI fencing first then dual PDU fencing if that fails In a node failure scenario, Pacemaker will first select ``fence_ipmilan`` to try to kill the faulty node. Using the fencing topology, if that method fails, it will then move on to selecting ``fence_apc_snmp`` twice (once for the first PDU, then again for the second PDU). The fence action is considered successful only if both PDUs report the required status. If any of them fails, fencing loops back to the first fencing method, ``fence_ipmilan``, and so on, until the node is fenced or the fencing action is cancelled. .. note:: **First fencing method: single IPMI device per target** Each cluster node has it own dedicated IPMI controller that can be contacted for fencing using the following primitives: .. code-block:: xml .. note:: **Second fencing method: dual PDU devices** Each cluster node also has 2 distinct power supplies controlled by 2 distinct PDUs: * Node 1: PDU 1 port 10 and PDU 2 port 10 * Node 2: PDU 1 port 11 and PDU 2 port 11 The matching fencing agents are configured as follows: .. code-block:: xml .. note:: **Fencing topology** Now that all the fencing resources are defined, it's time to create the right topology. We want to first fence using IPMI and if that does not work, fence both PDUs to effectively and surely kill the node. .. code-block:: xml In ``fencing-topology``, the lowest ``index`` value for a target determines its first fencing method. Remapping Reboots ################# When the cluster needs to reboot a node, whether because ``stonith-action`` is ``reboot`` or because a reboot was requested externally (such as by ``stonith_admin --reboot``), it will remap that to other commands in two cases: * If the chosen fencing device does not support the ``reboot`` command, the cluster will ask it to perform ``off`` instead. * If a fencing topology level with multiple devices must be executed, the cluster will ask all the devices to perform ``off``, then ask the devices to perform ``on``. To understand the second case, consider the example of a node with redundant power supplies connected to intelligent power switches. Rebooting one switch and then the other would have no effect on the node. Turning both switches off, and then on, actually reboots the node. In such a case, the fencing operation will be treated as successful as long as the ``off`` commands succeed, because then it is safe for the cluster to recover any resources that were on the node. Timeouts and errors in the ``on`` phase will be logged but ignored. When a reboot operation is remapped, any action-specific timeout for the remapped action will be used (for example, ``pcmk_off_timeout`` will be used when executing the ``off`` command, not ``pcmk_reboot_timeout``).