diff --git a/cts/CIB.py b/cts/CIB.py index 8b6f3056e5..755c258146 100644 --- a/cts/CIB.py +++ b/cts/CIB.py @@ -1,521 +1,521 @@ """ CIB generator for Pacemaker's Cluster Test Suite (CTS) """ # Pacemaker targets compatibility with Python 2.7 and 3.2+ from __future__ import print_function, unicode_literals, absolute_import, division -__copyright__ = "Copyright 2008-2019 the Pacemaker project contributors" +__copyright__ = "Copyright 2008-2020 the Pacemaker project contributors" __license__ = "GNU General Public License version 2 or later (GPLv2+) WITHOUT ANY WARRANTY" import os import warnings import tempfile from cts.CTSvars import * class CibBase(object): def __init__(self, Factory, tag, _id, **kwargs): self.tag = tag self.name = _id self.kwargs = kwargs self.children = [] self.Factory = Factory def __repr__(self): return "%s-%s" % (self.tag, self.name) def add_child(self, child): self.children.append(child) def __setitem__(self, key, value): if value: self.kwargs[key] = value else: self.kwargs.pop(key, None) from cts.cib_xml import * class ConfigBase(object): cts_cib = None version = "unknown" Factory = None def __init__(self, CM, factory, tmpfile=None): self.CM = CM self.Factory = factory if not tmpfile: warnings.filterwarnings("ignore") f=tempfile.NamedTemporaryFile(delete=True) f.close() tmpfile = f.name warnings.resetwarnings() self.Factory.tmpfile = tmpfile def version(self): return self.version def NextIP(self): ip = self.CM.Env["IPBase"] if ":" in ip: (prefix, sep, suffix) = ip.rpartition(":") suffix = str(hex(int(suffix, 16)+1)).lstrip("0x") else: (prefix, sep, suffix) = ip.rpartition(".") suffix = str(int(suffix)+1) ip = prefix + sep + suffix self.CM.Env["IPBase"] = ip return ip.strip() class CIB12(ConfigBase): version = "pacemaker-1.2" counter = 1 def _show(self, command=""): output = "" (rc, result) = self.Factory.rsh(self.Factory.target, "HOME=/root CIB_file="+self.Factory.tmpfile+" cibadmin -Ql "+command, None, ) for line in result: output += line self.Factory.debug("Generated Config: "+line) return output def NewIP(self, name=None, standard="ocf"): if self.CM.Env["IPagent"] == "IPaddr2": ip = self.NextIP() if not name: if ":" in ip: (prefix, sep, suffix) = ip.rpartition(":") name = "r"+suffix else: name = "r"+ip r = Resource(self.Factory, name, self.CM.Env["IPagent"], standard) r["ip"] = ip if ":" in ip: r["cidr_netmask"] = "64" r["nic"] = "eth0" else: r["cidr_netmask"] = "32" else: if not name: name = "r%s%d" % (self.CM.Env["IPagent"], self.counter) self.counter = self.counter + 1 r = Resource(self.Factory, name, self.CM.Env["IPagent"], standard) r.add_op("monitor", "5s") return r def get_node_id(self, node_name): """ Check the cluster configuration for a node ID. """ # We can't account for every possible configuration, # so we only return a node ID if: # * The node is specified in /etc/corosync/corosync.conf # with "ring0_addr:" equal to node_name and "nodeid:" # explicitly specified. # In all other cases, we return 0. node_id = 0 # awkward command: use } as record separator # so each corosync.conf "object" is one record; # match the "node {" record that has "ring0_addr: node_name"; # then print the substring of that record after "nodeid:" (rc, output) = self.Factory.rsh(self.Factory.target, r"""awk -v RS="}" """ r"""'/^(\s*nodelist\s*{)?\s*node\s*{.*(ring0_addr|name):\s*%s(\s+|$)/""" r"""{gsub(/.*nodeid:\s*/,"");gsub(/\s+.*$/,"");print}'""" r""" /etc/corosync/corosync.conf""" % node_name, None) if rc == 0 and len(output) == 1: try: node_id = int(output[0]) except ValueError: node_id = 0 return node_id def install(self, target): old = self.Factory.tmpfile # Force a rebuild self.cts_cib = None self.Factory.tmpfile = CTSvars.CRM_CONFIG_DIR+"/cib.xml" self.contents(target) self.Factory.rsh(self.Factory.target, "chown "+CTSvars.CRM_DAEMON_USER+" "+self.Factory.tmpfile) self.Factory.tmpfile = old def contents(self, target=None): # fencing resource if self.cts_cib: return self.cts_cib if target: self.Factory.target = target self.Factory.rsh(self.Factory.target, "HOME=/root cibadmin --empty %s > %s" % (self.version, self.Factory.tmpfile)) self.num_nodes = len(self.CM.Env["nodes"]) no_quorum = "stop" if self.num_nodes < 3: no_quorum = "ignore" self.Factory.log("Cluster only has %d nodes, configuring: no-quorum-policy=ignore" % self.num_nodes) # We don't need a nodes section unless we add attributes stn = None # Fencing resource # Define first so that the shell doesn't reject every update if self.CM.Env["DoFencing"]: # Define the "real" fencing device st = Resource(self.Factory, "Fencing", ""+self.CM.Env["stonith-type"], "stonith") # Set a threshold for unreliable stonith devices such as the vmware one st.add_meta("migration-threshold", "5") st.add_op("monitor", "120s", timeout="120s") st.add_op("stop", "0", timeout="60s") st.add_op("start", "0", timeout="60s") # For remote node tests, a cluster node is stopped and brought back up # as a remote node with the name "remote-OLDNAME". To allow fencing # devices to fence these nodes, create a list of all possible node names. all_node_names = [ prefix+n for n in self.CM.Env["nodes"] for prefix in ('', 'remote-') ] # Add all parameters specified by user entries = self.CM.Env["stonith-params"].split(',') for entry in entries: try: (name, value) = entry.split('=', 1) except ValueError: print("Warning: skipping invalid fencing parameter: %s" % entry) continue # Allow user to specify "all" as the node list, and expand it here if name in [ "hostlist", "pcmk_host_list" ] and value == "all": value = ' '.join(all_node_names) st[name] = value st.commit() # Test advanced fencing logic if True: stf_nodes = [] stt_nodes = [] attr_nodes = {} # Create the levels stl = FencingTopology(self.Factory) for node in self.CM.Env["nodes"]: # Remote node tests will rename the node remote_node = "remote-" + node # Randomly assign node to a fencing method ftype = self.CM.Env.RandomGen.choice(["levels-and", "levels-or ", "broadcast "]) # For levels-and, randomly choose targeting by node name or attribute by = "" if ftype == "levels-and": node_id = self.get_node_id(node) if node_id == 0 or self.CM.Env.RandomGen.choice([True, False]): by = " (by name)" else: attr_nodes[node] = node_id by = " (by attribute)" self.CM.log(" - Using %s fencing for node: %s%s" % (ftype, node, by)) if ftype == "levels-and": # If targeting by name, add a topology level for this node if node not in attr_nodes: stl.level(1, node, "FencingPass,Fencing") # Always target remote nodes by name, otherwise we would need to add # an attribute to the remote node only during remote tests (we don't # want nonexistent remote nodes showing up in the non-remote tests). # That complexity is not worth the effort. stl.level(1, remote_node, "FencingPass,Fencing") # Add the node (and its remote equivalent) to the list of levels-and nodes. stt_nodes.extend([node, remote_node]) elif ftype == "levels-or ": for n in [ node, remote_node ]: stl.level(1, n, "FencingFail") stl.level(2, n, "Fencing") stf_nodes.extend([node, remote_node]) # If any levels-and nodes were targeted by attribute, # create the attributes and a level for the attribute. if attr_nodes: stn = Nodes(self.Factory) for (node_name, node_id) in list(attr_nodes.items()): stn.add_node(node_name, node_id, { "cts-fencing" : "levels-and" }) stl.level(1, None, "FencingPass,Fencing", "cts-fencing", "levels-and") # Create a Dummy agent that always passes for levels-and if len(stt_nodes): stt = Resource(self.Factory, "FencingPass", "fence_dummy", "stonith") stt["pcmk_host_list"] = " ".join(stt_nodes) # Wait this many seconds before doing anything, handy for letting disks get flushed too stt["random_sleep_range"] = "30" stt["mode"] = "pass" stt.commit() # Create a Dummy agent that always fails for levels-or if len(stf_nodes): stf = Resource(self.Factory, "FencingFail", "fence_dummy", "stonith") stf["pcmk_host_list"] = " ".join(stf_nodes) # Wait this many seconds before doing anything, handy for letting disks get flushed too stf["random_sleep_range"] = "30" stf["mode"] = "fail" stf.commit() # Now commit the levels themselves stl.commit() o = Option(self.Factory) o["stonith-enabled"] = self.CM.Env["DoFencing"] o["start-failure-is-fatal"] = "false" o["pe-input-series-max"] = "5000" o["shutdown-escalation"] = "5min" o["batch-limit"] = "10" o["dc-deadtime"] = "5s" o["no-quorum-policy"] = no_quorum if self.CM.Env["DoBSC"] == 1: o["ident-string"] = "Linux-HA TEST configuration file - REMOVEME!!" o.commit() o = OpDefaults(self.Factory) o["timeout"] = "90s" o.commit() # Commit the nodes section if we defined one if stn is not None: stn.commit() # Add an alerts section if possible if self.Factory.rsh.exists_on_all(self.CM.Env["notification-agent"], self.CM.Env["nodes"]): alerts = Alerts(self.Factory) alerts.add_alert(self.CM.Env["notification-agent"], self.CM.Env["notification-recipient"]) alerts.commit() # Add resources? if self.CM.Env["CIBResource"] == 1: self.add_resources() if self.CM.cluster_monitor == 1: mon = Resource(self.Factory, "cluster_mon", "ocf", "ClusterMon", "pacemaker") mon.add_op("start", "0", requires="nothing") mon.add_op("monitor", "5s", requires="nothing") mon["update"] = "10" mon["extra_options"] = "-r -n" mon["user"] = "abeekhof" mon["htmlfile"] = "/suse/abeekhof/Export/cluster.html" mon.commit() #self._create('''location prefer-dc cluster_mon rule -INFINITY: \#is_dc eq false''') # generate cib self.cts_cib = self._show() if self.Factory.tmpfile != CTSvars.CRM_CONFIG_DIR+"/cib.xml": self.Factory.rsh(self.Factory.target, "rm -f "+self.Factory.tmpfile) return self.cts_cib def add_resources(self): # Per-node resources for node in self.CM.Env["nodes"]: name = "rsc_"+node r = self.NewIP(name) r.prefer(node, "100") r.commit() # Migrator # Make this slightly sticky (since we have no other location constraints) to avoid relocation during Reattach m = Resource(self.Factory, "migrator","Dummy", "ocf", "pacemaker") m["passwd"] = "whatever" m.add_meta("resource-stickiness","1") m.add_meta("allow-migrate", "1") m.add_op("monitor", "P10S") m.commit() - # Ping the test master + # Ping the test exerciser p = Resource(self.Factory, "ping-1","ping", "ocf", "pacemaker") p.add_op("monitor", "60s") - p["host_list"] = self.CM.Env["cts-master"] + p["host_list"] = self.CM.Env["cts-exerciser"] p["name"] = "connected" p["debug"] = "true" c = Clone(self.Factory, "Connectivity", p) c["globally-unique"] = "false" c.commit() # promotable clone resource s = Resource(self.Factory, "stateful-1", "Stateful", "ocf", "pacemaker") s.add_op("monitor", "15s", timeout="60s") s.add_op("monitor", "16s", timeout="60s", role="Master") ms = Clone(self.Factory, "promotable-1", s) ms["promotable"] = "true" ms["clone-max"] = self.num_nodes ms["clone-node-max"] = 1 ms["promoted-max"] = 1 ms["promoted-node-max"] = 1 # Require connectivity to run the promotable clone r = Rule(self.Factory, "connected", "-INFINITY", op="or") r.add_child(Expression(self.Factory, "m1-connected-1", "connected", "lt", "1")) r.add_child(Expression(self.Factory, "m1-connected-2", "connected", "not_defined", None)) ms.prefer("connected", rule=r) ms.commit() # Group Resource g = Group(self.Factory, "group-1") g.add_child(self.NewIP()) if self.CM.Env["have_systemd"]: sysd = Resource(self.Factory, "petulant", "pacemaker-cts-dummyd@10", "service") sysd.add_op("monitor", "P10S") g.add_child(sysd) else: g.add_child(self.NewIP()) g.add_child(self.NewIP()) # Make group depend on the promotable clone g.after("promotable-1", first="promote", then="start") g.colocate("promotable-1", "INFINITY", withrole="Master") g.commit() # LSB resource lsb = Resource(self.Factory, "lsb-dummy", "LSBDummy", "lsb") lsb.add_op("monitor", "5s") # LSB with group lsb.after("group-1") lsb.colocate("group-1") lsb.commit() class CIB20(CIB12): version = "pacemaker-2.5" class CIB30(CIB12): version = "pacemaker-3.0" #class HASI(CIB10): # def add_resources(self): # # DLM resource # self._create('''primitive dlm ocf:pacemaker:controld op monitor interval=120s''') # self._create('''clone dlm-clone dlm meta globally-unique=false interleave=true''') # O2CB resource # self._create('''primitive o2cb ocf:ocfs2:o2cb op monitor interval=120s''') # self._create('''clone o2cb-clone o2cb meta globally-unique=false interleave=true''') # self._create('''colocation o2cb-with-dlm INFINITY: o2cb-clone dlm-clone''') # self._create('''order start-o2cb-after-dlm mandatory: dlm-clone o2cb-clone''') class ConfigFactory(object): def __init__(self, CM): self.CM = CM self.rsh = self.CM.rsh self.register("pacemaker12", CIB12, CM, self) self.register("pacemaker20", CIB20, CM, self) self.register("pacemaker30", CIB30, CM, self) # self.register("hae", HASI, CM, self) if self.CM.Env["ListTests"] == 0: self.target = self.CM.Env["nodes"][0] self.tmpfile = None def log(self, args): self.CM.log("cib: %s" % args) def debug(self, args): self.CM.debug("cib: %s" % args) def register(self, methodName, constructor, *args, **kargs): """register a constructor""" _args = [constructor] _args.extend(args) setattr(self, methodName, ConfigFactoryItem(*_args, **kargs)) def unregister(self, methodName): """unregister a constructor""" delattr(self, methodName) def createConfig(self, name="pacemaker-1.0"): if name == "pacemaker-1.0": name = "pacemaker10"; elif name == "pacemaker-1.2": name = "pacemaker12"; elif name == "pacemaker-2.0": name = "pacemaker20"; elif name == "pacemaker-3.0": name = "pacemaker30"; elif name == "hasi": name = "hae"; if hasattr(self, name): return getattr(self, name)() else: self.CM.log("Configuration variant '%s' is unknown. Defaulting to latest config" % name) return self.pacemaker30() class ConfigFactoryItem(object): def __init__(self, function, *args, **kargs): self._function = function self._args = args self._kargs = kargs def __call__(self, *args, **kargs): """call function""" _args = list(self._args) _args.extend(args) _kargs = self._kargs.copy() _kargs.update(kargs) return self._function(*_args,**_kargs) if __name__ == '__main__': """ Unit test (pass cluster node names as command line arguments) """ import cts.CTS import cts.CM_corosync import sys if len(sys.argv) < 2: print("Usage: %s ..." % sys.argv[0]) sys.exit(1) args = [ "--nodes", " ".join(sys.argv[1:]), "--clobber-cib", "--populate-resources", "--stack", "corosync", "--test-ip-base", "fe80::1234:56:7890:1000", "--stonith", "rhcs", ] env = CTS.CtsLab(args) cm = CM_corosync.crm_corosync(env) CibFactory = ConfigFactory(cm) cib = CibFactory.createConfig("pacemaker-2.0") print(cib.contents()) diff --git a/cts/CTSlab.py.in b/cts/CTSlab.py.in index f4ae60dc13..4bae935151 100644 --- a/cts/CTSlab.py.in +++ b/cts/CTSlab.py.in @@ -1,130 +1,130 @@ #!@PYTHON@ """ Command-line interface to Pacemaker's Cluster Test Suite (CTS) """ # Pacemaker targets compatibility with Python 2.7 and 3.2+ from __future__ import print_function, unicode_literals, absolute_import, division -__copyright__ = "Copyright 2001-2018 the Pacemaker project contributors" +__copyright__ = "Copyright 2001-2020 the Pacemaker project contributors" __license__ = "GNU General Public License version 2 or later (GPLv2+) WITHOUT ANY WARRANTY" import sys, signal, os pdir = os.path.dirname(sys.path[0]) sys.path.insert(0, pdir) # So that things work from the source directory try: from cts.CTSvars import * from cts.CM_corosync import * from cts.CTSaudits import AuditList from cts.CTStests import TestList from cts.CTSscenarios import * from cts.logging import LogFactory except ImportError as e: sys.stderr.write("abort: %s\n" % e) sys.stderr.write("check your install and PYTHONPATH; couldn't find cts libraries in:\n%s\n" % ' '.join(sys.path)) sys.exit(1) # These are globals so they can be used by the signal handler. scenario = None LogFactory().add_stderr() def sig_handler(signum, frame) : LogFactory().log("Interrupted by signal %d"%signum) if scenario: scenario.summarize() if signum == 15 : if scenario: scenario.TearDown() sys.exit(1) if __name__ == '__main__': Environment = CtsLab(sys.argv[1:]) NumIter = Environment["iterations"] if NumIter is None: NumIter = 1 Tests = [] # Set the signal handler signal.signal(15, sig_handler) signal.signal(10, sig_handler) # Create the Cluster Manager object cm = None if Environment["Stack"] == "corosync 2+": cm = crm_corosync(Environment) else: LogFactory().log("Unknown stack: "+Environment["stack"]) sys.exit(1) if Environment["TruncateLog"] == 1: if Environment["OutputFile"] is None: LogFactory().log("Ignoring truncate request because no output file specified") else: LogFactory().log("Truncating %s" % Environment["OutputFile"]) with open(Environment["OutputFile"], "w") as outputfile: outputfile.truncate(0) Audits = AuditList(cm) if Environment["ListTests"] == 1: Tests = TestList(cm, Audits) LogFactory().log("Total %d tests"%len(Tests)) for test in Tests : LogFactory().log(str(test.name)); sys.exit(0) elif len(Environment["tests"]) == 0: Tests = TestList(cm, Audits) else: Chosen = Environment["tests"] for TestCase in Chosen: match = None for test in TestList(cm, Audits): if test.name == TestCase: match = test if not match: LogFactory().log("--choose: No applicable/valid tests chosen") sys.exit(1) else: Tests.append(match) # Scenario selection if Environment["scenario"] == "basic-sanity": scenario = RandomTests(cm, [ BasicSanityCheck(Environment) ], Audits, Tests) elif Environment["scenario"] == "all-once": NumIter = len(Tests) scenario = AllOnce( cm, [ BootCluster(Environment), PacketLoss(Environment) ], Audits, Tests) elif Environment["scenario"] == "sequence": scenario = Sequence( cm, [ BootCluster(Environment), PacketLoss(Environment) ], Audits, Tests) elif Environment["scenario"] == "boot": scenario = Boot(cm, [ LeaveBooted(Environment)], Audits, []) else: scenario = RandomTests( cm, [ BootCluster(Environment), PacketLoss(Environment) ], Audits, Tests) LogFactory().log(">>>>>>>>>>>>>>>> BEGINNING " + repr(NumIter) + " TESTS ") LogFactory().log("Stack: %s (%s)" % (Environment["Stack"], Environment["Name"])) LogFactory().log("Schema: %s" % Environment["Schema"]) LogFactory().log("Scenario: %s" % scenario.__doc__) - LogFactory().log("CTS Master: %s" % Environment["cts-master"]) + LogFactory().log("CTS Exerciser: %s" % Environment["cts-exerciser"]) LogFactory().log("CTS Logfile: %s" % Environment["OutputFile"]) LogFactory().log("Random Seed: %s" % Environment["RandSeed"]) LogFactory().log("Syslog variant: %s" % Environment["syslogd"].strip()) LogFactory().log("System log files: %s" % Environment["LogFileName"]) if Environment.has_key("IPBase"): LogFactory().log("Base IP for resources: %s" % Environment["IPBase"]) LogFactory().log("Cluster starts at boot: %d" % Environment["at-boot"]) Environment.dump() rc = Environment.run(scenario, NumIter) sys.exit(rc) diff --git a/cts/cluster_test.in b/cts/cluster_test.in index e0d28509d9..38f941d3ec 100755 --- a/cts/cluster_test.in +++ b/cts/cluster_test.in @@ -1,174 +1,175 @@ #!@BASH_PATH@ # -# Copyright 2008-2018 the Pacemaker project contributors +# Copyright 2008-2020 the Pacemaker project contributors # # The version control history for this file may have further details. # # This source code is licensed under the GNU General Public License version 2 # or later (GPLv2+) WITHOUT ANY WARRANTY. # if [ -e ~/.cts ]; then . ~/.cts fi anyAsked=0 [ $# -lt 1 ] || CTS_numtests=$1 die() { echo "$@"; exit 1; } if [ -z "$CTS_asked_once" ]; then anyAsked=1 - echo "This script should only be executed on the test master." - echo "The test master will remotely execute the actions required by the tests and should not be part of the cluster itself." + echo "This script should only be executed on the test exerciser." + echo "The test exerciser will remotely execute the actions required by the" + echo "tests and should not be part of the cluster itself." - read -p "Is this host intended to be the test master? (yN) " doUnderstand + read -p "Is this host intended to be the test exerciser? (yN) " doUnderstand [ "$doUnderstand" = "y" ] \ - || die "This script must be executed on the test master" + || die "This script must be executed on the test exerciser" fi if [ -z "$CTS_node_list" ]; then anyAsked=1 read -p "Please list your cluster nodes (eg. node1 node2 node3): " CTS_node_list else echo "Beginning test of cluster: $CTS_node_list" fi if [ -z "$CTS_stack" ]; then anyAsked=1 read -p "Which cluster stack are you using? ([corosync]): " CTS_stack [ -n "$CTS_stack" ] || CTS_stack=corosync else echo "Using the $CTS_stack cluster stack" fi [ "${CTS_node_list}" = "${CTS_node_list/$HOSTNAME/}" ] \ - || die "This script must be executed on the test master and the test master cannot be part of the cluster" + || die "This script must be executed on the test exerciser, and the test exerciser cannot be part of the cluster" printf "+ Bootstraping ssh... " if [ -z "$SSH_AUTH_SOCK" ]; then printf "\n + Initializing SSH " eval "$(ssh-agent)" echo " + Adding identities..." ssh-add rc=$? if [ $rc -ne 0 ]; then echo " -- No identities added" printf "\nThe ability to open key-based 'ssh' connections (as the user 'root') is required to use CTS.\n" read -p " - Do you want this program to help you create one? (yN) " auto_fix if [ "$auto_fix" = "y" ]; then ssh-keygen -t dsa ssh-add else die "Please run 'ssh-keygen -t dsa' to create a new key" fi fi else echo "OK" fi test_ok=1 printf "+ Testing ssh configuration... " for n in $CTS_node_list; do ssh -l root -o PasswordAuthentication=no -o ConnectTimeout=5 "$n" /bin/true rc=$? if [ $rc -ne 0 ]; then echo " - connection to $n failed" test_ok=0 fi done if [ $test_ok -eq 0 ]; then printf "\nThe ability to open key-based 'ssh' connections (as the user 'root') is required to use CTS.\n" read -p " - Do you want this program to help you with such a setup? (yN) " auto_fix if [ "$auto_fix" = "y" ]; then # XXX are we picking the most suitable identity? privKey=$(ssh-add -L | head -n1 | cut -d" " -f3) sshCopyIdOpts="-o User=root" [ -z "$privKey" ] || sshCopyIdOpts+=" -i \"${privKey}.pub\"" for n in $CTS_node_list; do eval "ssh-copy-id $sshCopyIdOpts \"${n}\"" \ || die "Attempt to 'ssh-copy-id $sshCopyIdOpts \"$n\"' failed" done else die "Please install one of your SSH public keys to root's account on all cluster nodes" fi fi echo "OK" if [ -z "$CTS_logfile" ]; then anyAsked=1 read -p " + Where does/should syslog store logs from remote hosts? (/var/log/messages) " CTS_logfile [ -n "$CTS_logfile" ] || CTS_logfile=/var/log/messages fi [ -e "$CTS_logfile" ] || die "$CTS_logfile doesn't exist" if [ -z "$CTS_logfacility" ]; then anyAsked=1 read -p " + Which log facility does the cluster use? (daemon) " CTS_logfacility [ -n "$CTS_logfacility" ] || CTS_logfacility=daemon fi if [ -z "$CTS_boot" ]; then read -p "+ Is the cluster software started automatically when a node boots? [yN] " CTS_boot if [ -z "$CTS_boot" ]; then CTS_boot=0 else case $CTS_boot in 1|y|Y) CTS_boot=1;; *) CTS_boot=0;; esac fi fi if [ -z "$CTS_numtests" ]; then read -p "+ How many test iterations should be performed? (500) " CTS_numtests [ -n "$CTS_numtests" ] || CTS_numtests=500 fi if [ -z "$CTS_asked_once" ]; then anyAsked=1 read -p "+ What type of STONITH agent do you use? (none) " CTS_stonith [ -z "$CTS_stonith" ] \ || read -p "+ List any STONITH agent parameters (eq. device_host=switch.power.com): " CTS_stonith_args [ -n "$CTS_adv" ] \ || read -p "+ (Advanced) Any extra CTS parameters? (none) " CTS_adv fi [ $anyAsked -eq 0 ] \ || read -p "+ Save values to ~/.cts for next time? (yN) " doSave if [ "$doSave" = "y" ]; then cat > ~/.cts <<-EOF # CTS Test data CTS_stack="$CTS_stack" CTS_node_list="$CTS_node_list" CTS_logfile="$CTS_logfile" CTS_logport="$CTS_logport" CTS_logfacility="$CTS_logfacility" CTS_asked_once=1 CTS_adv="$CTS_adv" CTS_stonith="$CTS_stonith" CTS_stonith_args="$CTS_stonith_args" CTS_boot="$CTS_boot" EOF fi cts_extra="" if [ -n "$CTS_stonith" ]; then cts_extra="$cts_extra --stonith-type $CTS_stonith" [ -z "$CTS_stonith_args" ] \ || cts_extra="$cts_extra --stonith-params \"$CTS_stonith_args\"" else cts_extra="$cts_extra --stonith 0" echo " - Testing a cluster without STONITH is like a blunt pencil... pointless" fi printf "\nAll set to go for %d iterations!\n" "$CTS_numtests" [ $anyAsked -ne 0 ] \ || echo "+ To use a different configuration, remove ~/.cts and re-run cts (or edit it manually)." echo Now paste the following command into this shell: echo "@PYTHON@ `dirname "$0"`/CTSlab.py -L \"$CTS_logfile\" --syslog-facility \"$CTS_logfacility\" --no-unsafe-tests --stack \"$CTS_stack\" $CTS_adv --at-boot \"$CTS_boot\" $cts_extra \"$CTS_numtests\" --nodes \"$CTS_node_list\"" diff --git a/cts/environment.py b/cts/environment.py index db9d3db164..6a97b12899 100644 --- a/cts/environment.py +++ b/cts/environment.py @@ -1,655 +1,655 @@ """ Test environment classes for Pacemaker's Cluster Test Suite (CTS) """ # Pacemaker targets compatibility with Python 2.7 and 3.2+ from __future__ import print_function, unicode_literals, absolute_import, division __copyright__ = "Copyright 2014-2019 the Pacemaker project contributors" __license__ = "GNU General Public License version 2 or later (GPLv2+) WITHOUT ANY WARRANTY" import sys, time, os, socket, random from cts.remote import * from cts.CTSvars import * class Environment(object): def __init__(self, args): self.data = {} self.Nodes = [] self["DeadTime"] = 300 self["StartTime"] = 300 self["StableTime"] = 30 self["tests"] = [] self["IPagent"] = "IPaddr2" self["DoStandby"] = 1 self["DoFencing"] = 1 self["XmitLoss"] = "0.0" self["RecvLoss"] = "0.0" self["ClobberCIB"] = 0 self["CIBfilename"] = None self["CIBResource"] = 0 self["DoBSC"] = 0 self["oprofile"] = [] self["warn-inactive"] = 0 self["ListTests"] = 0 self["benchmark"] = 0 self["LogWatcher"] = "any" self["SyslogFacility"] = "daemon" self["LogFileName"] = "/var/log/messages" self["Schema"] = "pacemaker-3.0" self["Stack"] = "corosync" self["stonith-type"] = "external/ssh" self["stonith-params"] = "hostlist=all,livedangerously=yes" self["notification-agent"] = "/var/lib/pacemaker/notify.sh" self["notification-recipient"] = "/var/lib/pacemaker/notify.log" self["loop-minutes"] = 60 self["valgrind-procs"] = "pacemaker-attrd pacemaker-based pacemaker-controld pacemaker-execd pacemaker-fenced pacemaker-schedulerd" self["experimental-tests"] = 0 self["container-tests"] = 0 self["valgrind-tests"] = 0 self["unsafe-tests"] = 1 self["loop-tests"] = 1 self["scenario"] = "random" self["stats"] = 0 self["docker"] = 0 self["continue"] = 0 self.RandomGen = random.Random() self.logger = LogFactory() self.SeedRandom() self.rsh = RemoteFactory().getInstance() self.target = "localhost" self.parse_args(args) if self["ListTests"] == 0: self.validate() self.discover() def SeedRandom(self, seed=None): if not seed: seed = int(time.time()) self["RandSeed"] = seed self.RandomGen.seed(str(seed)) def dump(self): keys = [] for key in list(self.data.keys()): keys.append(key) keys.sort() for key in keys: self.logger.debug("Environment["+key+"]:\t"+str(self[key])) def keys(self): return list(self.data.keys()) def has_key(self, key): if key == "nodes": return True return key in self.data def __getitem__(self, key): if str(key) == "0": raise ValueError("Bad call to 'foo in X', should reference 'foo in X.keys()' instead") if key == "nodes": return self.Nodes elif key == "Name": return self.get_stack_short() elif key in self.data: return self.data[key] else: return None def __setitem__(self, key, value): if key == "Stack": self.set_stack(value) elif key == "node-limit": self.data[key] = value self.filter_nodes() elif key == "nodes": self.Nodes = [] for node in value: # I don't think I need the IP address, etc. but this validates # the node name against /etc/hosts and/or DNS, so it's a # GoodThing(tm). try: n = node.strip() if self.data["docker"] == 0: socket.gethostbyname_ex(n) self.Nodes.append(n) except: self.logger.log(node+" not found in DNS... aborting") raise self.filter_nodes() else: self.data[key] = value def RandomNode(self): '''Choose a random node from the cluster''' return self.RandomGen.choice(self["nodes"]) def set_stack(self, name): # Normalize stack names if name == "corosync" or name == "cs" or name == "mcp": self.data["Stack"] = "corosync 2+" else: raise ValueError("Unknown stack: "+name) def get_stack_short(self): # Create the Cluster Manager object if not "Stack" in self.data: return "unknown" elif self.data["Stack"] == "corosync 2+": if self["docker"]: return "crm-corosync-docker" else: return "crm-corosync" else: LogFactory().log("Unknown stack: "+self["stack"]) raise ValueError("Unknown stack: "+self["stack"]) def detect_syslog(self): # Detect syslog variant if not "syslogd" in self.data: if self["have_systemd"]: # Systemd self["syslogd"] = self.rsh(self.target, "systemctl list-units | grep syslog.*\.service.*active.*running | sed 's:.service.*::'", stdout=1).strip() else: # SYS-V self["syslogd"] = self.rsh(self.target, "chkconfig --list | grep syslog.*on | awk '{print $1}' | head -n 1", stdout=1).strip() if not "syslogd" in self.data or not self["syslogd"]: # default self["syslogd"] = "rsyslog" def disable_service(self, node, service): if self["have_systemd"]: # Systemd return self.rsh(node, "systemctl disable %s" % service) else: # SYS-V return self.rsh(node, "chkconfig %s off" % service) def enable_service(self, node, service): if self["have_systemd"]: # Systemd return self.rsh(node, "systemctl enable %s" % service) else: # SYS-V return self.rsh(node, "chkconfig %s on" % service) def service_is_enabled(self, node, service): if self["have_systemd"]: # Systemd # With "systemctl is-enabled", we should check if the service is # explicitly "enabled" instead of the return code. For example it returns # 0 if the service is "static" or "indirect", but they don't really count # as "enabled". return not self.rsh(node, "systemctl is-enabled %s | grep enabled" % service) else: # SYS-V return not self.rsh(node, "chkconfig --list | grep -e %s.*on" % service) def detect_at_boot(self): # Detect if the cluster starts at boot if not "at-boot" in self.data: self["at-boot"] = self.service_is_enabled(self.target, "corosync") \ or self.service_is_enabled(self.target, "pacemaker") def detect_ip_offset(self): # Try to determine an offset for IPaddr resources if self["CIBResource"] and not "IPBase" in self.data: network=self.rsh(self.target, "ip addr | grep inet | grep -v -e link -e inet6 -e '/32' -e ' lo' | awk '{print $2}'", stdout=1).strip() self["IPBase"] = self.rsh(self.target, "nmap -sn -n %s | grep 'scan report' | awk '{print $NF}' | sed 's:(::' | sed 's:)::' | sort -V | tail -n 1" % network, stdout=1).strip() if not self["IPBase"]: self["IPBase"] = " fe80::1234:56:7890:1000" self.logger.log("Could not determine an offset for IPaddr resources. Perhaps nmap is not installed on the nodes.") self.logger.log("Defaulting to '%s', use --test-ip-base to override" % self["IPBase"]) elif int(self["IPBase"].split('.')[3]) >= 240: self.logger.log("Could not determine an offset for IPaddr resources. Upper bound is too high: %s %s" % (self["IPBase"], self["IPBase"].split('.')[3])) self["IPBase"] = " fe80::1234:56:7890:1000" self.logger.log("Defaulting to '%s', use --test-ip-base to override" % self["IPBase"]) def filter_nodes(self): if self['node-limit'] is not None and self["node-limit"] > 0: if len(self["nodes"]) > self["node-limit"]: self.logger.log("Limiting the number of nodes configured=%d (max=%d)" %(len(self["nodes"]), self["node-limit"])) while len(self["nodes"]) > self["node-limit"]: self["nodes"].pop(len(self["nodes"])-1) def validate(self): if len(self["nodes"]) < 1: print("No nodes specified!") sys.exit(1) def discover(self): self.target = random.Random().choice(self["nodes"]) - master = socket.gethostname() + exerciser = socket.gethostname() # Use the IP where possible to avoid name lookup failures - for ip in socket.gethostbyname_ex(master)[2]: + for ip in socket.gethostbyname_ex(exerciser)[2]: if ip != "127.0.0.1": - master = ip + exerciser = ip break; - self["cts-master"] = master + self["cts-exerciser"] = exerciser if not "have_systemd" in self.data: self["have_systemd"] = not self.rsh(self.target, "systemctl list-units", silent=True) self.detect_syslog() self.detect_at_boot() self.detect_ip_offset() def parse_args(self, args): skipthis=None if not args: args=sys.argv[1:] for i in range(0, len(args)): if skipthis: skipthis=None continue elif args[i] == "-l" or args[i] == "--limit-nodes": skipthis=1 self["node-limit"] = int(args[i+1]) elif args[i] == "-r" or args[i] == "--populate-resources": self["CIBResource"] = 1 self["ClobberCIB"] = 1 elif args[i] == "--outputfile": skipthis=1 self["OutputFile"] = args[i+1] LogFactory().add_file(self["OutputFile"]) elif args[i] == "-L" or args[i] == "--logfile": skipthis=1 self["LogWatcher"] = "remote" self["LogAuditDisabled"] = 1 self["LogFileName"] = args[i+1] elif args[i] == "--ip" or args[i] == "--test-ip-base": skipthis=1 self["IPBase"] = args[i+1] self["CIBResource"] = 1 self["ClobberCIB"] = 1 elif args[i] == "--oprofile": skipthis=1 self["oprofile"] = args[i+1].split(' ') elif args[i] == "--trunc": self["TruncateLog"]=1 elif args[i] == "--list-tests" or args[i] == "--list" : self["ListTests"]=1 elif args[i] == "--benchmark": self["benchmark"]=1 elif args[i] == "--bsc": self["DoBSC"] = 1 self["scenario"] = "basic-sanity" elif args[i] == "--qarsh": RemoteFactory().enable_qarsh() elif args[i] == "--docker": self["docker"] = 1 RemoteFactory().enable_docker() elif args[i] == "--yes" or args[i] == "-y": self["continue"] = 1 elif args[i] == "--stonith" or args[i] == "--fencing": skipthis=1 if args[i+1] == "1" or args[i+1] == "yes": self["DoFencing"]=1 elif args[i+1] == "0" or args[i+1] == "no": self["DoFencing"]=0 elif args[i+1] == "phd": self["DoStonith"]=1 self["stonith-type"] = "fence_phd_kvm" elif args[i+1] == "rhcs" or args[i+1] == "xvm" or args[i+1] == "virt": self["DoStonith"]=1 self["stonith-type"] = "fence_xvm" elif args[i+1] == "docker": self["DoStonith"]=1 self["stonith-type"] = "fence_docker_cts" elif args[i+1] == "scsi": self["DoStonith"]=1 self["stonith-type"] = "fence_scsi" elif args[i+1] == "ssh" or args[i+1] == "lha": self["DoStonith"]=1 self["stonith-type"] = "external/ssh" self["stonith-params"] = "hostlist=all,livedangerously=yes" elif args[i+1] == "north": self["DoStonith"]=1 self["stonith-type"] = "fence_apc" self["stonith-params"] = "ipaddr=north-apc,login=apc,passwd=apc,pcmk_host_map=north-01:2;north-02:3;north-03:4;north-04:5;north-05:6;north-06:7;north-07:9;north-08:10;north-09:11;north-10:12;north-11:13;north-12:14;north-13:15;north-14:18;north-15:17;north-16:19;" elif args[i+1] == "south": self["DoStonith"]=1 self["stonith-type"] = "fence_apc" self["stonith-params"] = "ipaddr=south-apc,login=apc,passwd=apc,pcmk_host_map=south-01:2;south-02:3;south-03:4;south-04:5;south-05:6;south-06:7;south-07:9;south-08:10;south-09:11;south-10:12;south-11:13;south-12:14;south-13:15;south-14:18;south-15:17;south-16:19;" elif args[i+1] == "east": self["DoStonith"]=1 self["stonith-type"] = "fence_apc" self["stonith-params"] = "ipaddr=east-apc,login=apc,passwd=apc,pcmk_host_map=east-01:2;east-02:3;east-03:4;east-04:5;east-05:6;east-06:7;east-07:9;east-08:10;east-09:11;east-10:12;east-11:13;east-12:14;east-13:15;east-14:18;east-15:17;east-16:19;" elif args[i+1] == "west": self["DoStonith"]=1 self["stonith-type"] = "fence_apc" self["stonith-params"] = "ipaddr=west-apc,login=apc,passwd=apc,pcmk_host_map=west-01:2;west-02:3;west-03:4;west-04:5;west-05:6;west-06:7;west-07:9;west-08:10;west-09:11;west-10:12;west-11:13;west-12:14;west-13:15;west-14:18;west-15:17;west-16:19;" elif args[i+1] == "openstack": self["DoStonith"]=1 self["stonith-type"] = "fence_openstack" print("Obtaining OpenStack credentials from the current environment") self["stonith-params"] = "region=%s,tenant=%s,auth=%s,user=%s,password=%s" % ( os.environ['OS_REGION_NAME'], os.environ['OS_TENANT_NAME'], os.environ['OS_AUTH_URL'], os.environ['OS_USERNAME'], os.environ['OS_PASSWORD'] ) elif args[i+1] == "rhevm": self["DoStonith"]=1 self["stonith-type"] = "fence_rhevm" print("Obtaining RHEV-M credentials from the current environment") self["stonith-params"] = "login=%s,passwd=%s,ipaddr=%s,ipport=%s,ssl=1,shell_timeout=10" % ( os.environ['RHEVM_USERNAME'], os.environ['RHEVM_PASSWORD'], os.environ['RHEVM_SERVER'], os.environ['RHEVM_PORT'], ) else: self.usage(args[i+1]) elif args[i] == "--stonith-type": self["stonith-type"] = args[i+1] skipthis=1 elif args[i] == "--stonith-args": self["stonith-params"] = args[i+1] skipthis=1 elif args[i] == "--standby": skipthis=1 if args[i+1] == "1" or args[i+1] == "yes": self["DoStandby"] = 1 elif args[i+1] == "0" or args[i+1] == "no": self["DoStandby"] = 0 else: self.usage(args[i+1]) elif args[i] == "--clobber-cib" or args[i] == "-c": self["ClobberCIB"] = 1 elif args[i] == "--cib-filename": skipthis=1 self["CIBfilename"] = args[i+1] elif args[i] == "--xmit-loss": try: float(args[i+1]) except ValueError: print("--xmit-loss parameter should be float") self.usage(args[i+1]) skipthis=1 self["XmitLoss"] = args[i+1] elif args[i] == "--recv-loss": try: float(args[i+1]) except ValueError: print("--recv-loss parameter should be float") self.usage(args[i+1]) skipthis=1 self["RecvLoss"] = args[i+1] elif args[i] == "--choose": skipthis=1 self["tests"].append(args[i+1]) self["scenario"] = "sequence" elif args[i] == "--nodes": skipthis=1 self["nodes"] = args[i+1].split(' ') elif args[i] == "-g" or args[i] == "--group" or args[i] == "--dsh-group": skipthis=1 self["OutputFile"] = "%s/cluster-%s.log" % (os.environ['HOME'], args[i+1]) LogFactory().add_file(self["OutputFile"], "CTS") dsh_file = "%s/.dsh/group/%s" % (os.environ['HOME'], args[i+1]) # Hacks to make my life easier if args[i+1] == "virt1": self["Stack"] = "corosync" self["DoStonith"]=1 self["stonith-type"] = "fence_xvm" self["stonith-params"] = "delay=0" self["IPBase"] = " fe80::1234:56:7890:1000" elif args[i+1] == "east16" or args[i+1] == "nsew": self["Stack"] = "corosync" self["DoStonith"]=1 self["stonith-type"] = "fence_apc" self["stonith-params"] = "ipaddr=east-apc,login=apc,passwd=apc,pcmk_host_map=east-01:2;east-02:3;east-03:4;east-04:5;east-05:6;east-06:7;east-07:9;east-08:10;east-09:11;east-10:12;east-11:13;east-12:14;east-13:15;east-14:18;east-15:17;east-16:19;" self["IPBase"] = " fe80::1234:56:7890:2000" if args[i+1] == "east16": # Requires newer python than available via nsew self["IPagent"] = "Dummy" elif args[i+1] == "corosync8": self["Stack"] = "corosync" self["DoStonith"]=1 self["stonith-type"] = "fence_rhevm" print("Obtaining RHEV-M credentials from the current environment") self["stonith-params"] = "login=%s,passwd=%s,ipaddr=%s,ipport=%s,ssl=1,shell_timeout=10" % ( os.environ['RHEVM_USERNAME'], os.environ['RHEVM_PASSWORD'], os.environ['RHEVM_SERVER'], os.environ['RHEVM_PORT'], ) self["IPBase"] = " fe80::1234:56:7890:3000" if os.path.isfile(dsh_file): self["nodes"] = [] f = open(dsh_file, 'r') for line in f: l = line.strip().rstrip() if not l.startswith('#'): self["nodes"].append(l) f.close() else: print("Unknown DSH group: %s" % args[i+1]) elif args[i] == "--syslog-facility" or args[i] == "--facility": skipthis=1 self["SyslogFacility"] = args[i+1] elif args[i] == "--seed": skipthis=1 self.SeedRandom(args[i+1]) elif args[i] == "--warn-inactive": self["warn-inactive"] = 1 elif args[i] == "--schema": skipthis=1 self["Schema"] = args[i+1] elif args[i] == "--at-boot" or args[i] == "--cluster-starts-at-boot": skipthis=1 if args[i+1] == "1" or args[i+1] == "yes": self["at-boot"] = 1 elif args[i+1] == "0" or args[i+1] == "no": self["at-boot"] = 0 else: self.usage(args[i+1]) elif args[i] == "--stack": if args[i+1] == "fedora" or args[i+1] == "fedora-17" or args[i+1] == "fedora-18": self["Stack"] = "corosync" elif args[i+1] == "rhel-7": self["Stack"] = "corosync" else: self["Stack"] = args[i+1] skipthis=1 elif args[i] == "--once": self["scenario"] = "all-once" elif args[i] == "--boot": self["scenario"] = "boot" elif args[i] == "--notification-agent": self["notification-agent"] = args[i+1] skipthis = 1 elif args[i] == "--notification-recipient": self["notification-recipient"] = args[i+1] skipthis = 1 elif args[i] == "--valgrind-tests": self["valgrind-tests"] = 1 elif args[i] == "--valgrind-procs": self["valgrind-procs"] = args[i+1] skipthis = 1 elif args[i] == "--no-loop-tests": self["loop-tests"] = 0 elif args[i] == "--loop-minutes": skipthis=1 try: self["loop-minutes"]=int(args[i+1]) except ValueError: self.usage(args[i]) elif args[i] == "--no-unsafe-tests": self["unsafe-tests"] = 0 elif args[i] == "--experimental-tests": self["experimental-tests"] = 1 elif args[i] == "--container-tests": self["container-tests"] = 1 elif args[i] == "--set": skipthis=1 (name, value) = args[i+1].split('=') self[name] = value print("Setting %s = %s" % (name, value)) elif args[i] == "--help": self.usage(args[i], 0) elif args[i] == "--": break else: try: NumIter=int(args[i]) self["iterations"] = NumIter except ValueError: self.usage(args[i]) def usage(self, arg, status=1): if status: print("Illegal argument %s" % arg) print("usage: " + sys.argv[0] +" [options] number-of-iterations") print("\nCommon options: ") print("\t [--nodes 'node list'] list of cluster nodes separated by whitespace") print("\t [--group | -g 'name'] use the nodes listed in the named DSH group (~/.dsh/groups/$name)") print("\t [--limit-nodes max] only use the first 'max' cluster nodes supplied with --nodes") print("\t [--stack corosync] which cluster stack is installed") print("\t [--list-tests] list the valid tests") print("\t [--benchmark] add the timing information") print("\t ") print("Options that CTS will usually auto-detect correctly: ") print("\t [--logfile path] where should the test software look for logs from cluster nodes") print("\t [--syslog-facility name] which syslog facility should the test software log to") print("\t [--at-boot (1|0)] does the cluster software start at boot time") print("\t [--test-ip-base ip] offset for generated IP address resources") print("\t ") print("Options for release testing: ") print("\t [--populate-resources | -r] generate a sample configuration") print("\t [--choose name] run only the named test") print("\t [--stonith (1 | 0 | yes | no | rhcs | ssh)]") print("\t [--once] run all valid tests once") print("\t ") print("Additional (less common) options: ") print("\t [--clobber-cib | -c ] erase any existing configuration") print("\t [--outputfile path] optional location for the test software to write logs to") print("\t [--trunc] truncate logfile before starting") print("\t [--xmit-loss lost-rate(0.0-1.0)]") print("\t [--recv-loss lost-rate(0.0-1.0)]") print("\t [--standby (1 | 0 | yes | no)]") print("\t [--fencing (1 | 0 | yes | no | rhcs | lha | openstack )]") print("\t [--stonith-type type]") print("\t [--stonith-args name=value]") print("\t [--bsc]") print("\t [--notification-agent path] script to configure for Pacemaker alerts") print("\t [--notification-recipient r] recipient to pass to alert script") print("\t [--no-loop-tests] don't run looping/time-based tests") print("\t [--no-unsafe-tests] don't run tests that are unsafe for use with ocfs2/drbd") print("\t [--valgrind-tests] include tests using valgrind") print("\t [--experimental-tests] include experimental tests") print("\t [--container-tests] include pacemaker_remote tests that run in lxc container resources") print("\t [--oprofile 'node list'] list of cluster nodes to run oprofile on]") print("\t [--qarsh] use the QARSH backdoor to access nodes instead of SSH") print("\t [--docker] Indicates nodes are docker nodes.") print("\t [--seed random_seed]") print("\t [--set option=value]") print("\t [--yes | -y] continue to run cts when there is an interaction whether to continue running pacemaker-cts") print("\t ") print("\t Example: ") # @PYTHON@ would be better here but not worth making file this a .in print("\t python sys.argv[0] -g virt1 -r --stonith ssh --schema pacemaker-2.0 500") sys.exit(status) class EnvFactory(object): instance = None def __init__(self): pass def getInstance(self, args=None): if not EnvFactory.instance: EnvFactory.instance = Environment(args) return EnvFactory.instance