diff --git a/doc/Pacemaker_Administration/en-US/Ch-Tools.txt b/doc/Pacemaker_Administration/en-US/Ch-Tools.txt index f1728bea7b..661dba83be 100644 --- a/doc/Pacemaker_Administration/en-US/Ch-Tools.txt +++ b/doc/Pacemaker_Administration/en-US/Ch-Tools.txt @@ -1,464 +1,539 @@ :compat-mode: legacy = Using Pacemaker Command-Line Tools = +[[s-cmdline-output]] +== Controlling Command Line Output == +Some of the pacemaker command line utilities have been converted to a new output +system. Among these tools are `crm_mon` and `stonith_admin`. This is an ongoing +project, and more tools will be converted over time. This system lets you +control the formatting of output with `--output-as=` and the destination of output +with `--output-to=`. + +The available formats vary by tool, but at least plain text, HTML, and XML are +supported by all tools. The default format is plain text. The default +destination is stdout but can be redirected to any file. Some formats support +command line options for changing the style of the output. For instance: + +====== +------- +# crm_mon --help-output-html +Usage: + crm_mon [OPTION?] + +Provides a summary of cluster's current state. + +Outputs varying levels of detail in a number of different formats. + +Output Options (html): + --output-cgi Add text needed to use output in a CGI program + --output-meta-refresh=SECONDS How often to refresh + --output-stylesheet-link=URI Link to an external CSS stylesheet + --output-title=TITLE Page title +------- +====== + [[s-crm_mon]] == Monitor a Cluster with crm_mon == indexterm:[Command-line tool,crm_mon] The `crm_mon` utility displays the current state of an active cluster. It can show the cluster status organized by node or by resource, and can be used in either single-shot or dynamically updating mode. It can also display operations performed and information about failures. Using this tool, you can examine the state of the cluster for irregularities, and see how it responds when you cause or simulate failures. See the manual page or the output of `crm_mon --help` for a full description of its many options. .Sample output from crm_mon -1 ====== ------- -Stack: corosync -Current DC: node2 (version 2.0.0-1) - partition with quorum -Last updated: Mon Jan 29 12:18:42 2018 -Last change: Mon Jan 29 12:18:40 2018 by root via crm_attribute on node3 - -5 nodes configured -2 resources configured - -Online: [ node1 node2 node3 node4 node5 ] - -Active resources: - -Fencing (stonith:fence_xvm): Started node1 -IP (ocf:heartbeat:IPaddr2): Started node2 +Cluster Summary: + * Stack: corosync + * Current DC: node2 (version 2.0.0-1) - partition with quorum + * Last updated: Mon Jan 29 12:18:42 2018 + * Last change: Mon Jan 29 12:18:40 2018 by root via crm_attribute on node3 + * 5 nodes configured + * 2 resources configured + +Node List: + * Online: [ node1 node2 node3 node4 node5 ] + +* Active resources: + * Fencing (stonith:fence_xvm): Started node1 + * IP (ocf:heartbeat:IPaddr2): Started node2 ------- ====== .Sample output from crm_mon -n -1 ====== ------- -Stack: corosync -Current DC: node2 (version 2.0.0-1) - partition with quorum -Last updated: Mon Jan 29 12:21:48 2018 -Last change: Mon Jan 29 12:18:40 2018 by root via crm_attribute on node3 - -5 nodes configured -2 resources configured - -Node node1: online - Fencing (stonith:fence_xvm): Started -Node node2: online - IP (ocf:heartbeat:IPaddr2): Started -Node node3: online -Node node4: online -Node node5: online +Cluster Summary: + * Stack: corosync + * Current DC: node2 (version 2.0.0-1) - partition with quorum + * Last updated: Mon Jan 29 12:21:48 2018 + * Last change: Mon Jan 29 12:18:40 2018 by root via crm_attribute on node3 + * 5 nodes configured + * 2 resources configured + +* Node List: + * Node node1: online + * Fencing (stonith:fence_xvm): Started + * Node node2: online + * IP (ocf:heartbeat:IPaddr2): Started + * Node node3: online + * Node node4: online + * Node node5: online ------- ====== As mentioned in an earlier chapter, the DC is the node is where decisions are made. The cluster elects a node to be DC as needed. The only significance of the choice of DC to an administrator is the fact that its logs will have the most information about why decisions were made. +[[s-crm_mon-css]] +=== Styling crm_mon output === +indexterm:[Command-line tool,crm_mon,css] + +Various parts of `crm_mon`'s HTML output have a CSS class associated with them. +Not everything does, but some of the most interesting portions do. In the following +example, the status of each node has an "online" class and the details of each +resource have an "rsc-ok" class. + +====== +------- +

Node List

+ +------- +====== + +By default, a stylesheet for styling these classes is included in the head of +the HTML output. The relevant portions of this stylesheet that would be used +in the above example is: + +====== +------- + +------- +====== + +If you want to override some or all of the styling, simply create your own +stylesheet, place it on a web server, and pass `--output-stylesheet-link=` +to `crm_mon`. The link is added after the default stylesheet, so your +changes take precedence. You don't need to duplicate the entire default. +Only include what you want to change. [[s-cibadmin]] == Edit the CIB XML with cibadmin == indexterm:[Command-line tool,cibadmin] The most flexible tool for modifying the configuration is Pacemaker's `cibadmin` command. With `cibadmin`, you can query, add, remove, update or replace any part of the configuration. All changes take effect immediately, so there is no need to perform a reload-like operation. The simplest way of using `cibadmin` is to use it to save the current configuration to a temporary file, edit that file with your favorite text or XML editor, and then upload the revised configuration. .Safely using an editor to modify the cluster configuration ====== -------- # cibadmin --query > tmp.xml # vi tmp.xml # cibadmin --replace --xml-file tmp.xml -------- ====== Some of the better XML editors can make use of a RELAX NG schema to help make sure any changes you make are valid. The schema describing the configuration can be found in +pacemaker.rng+, which may be deployed in a location such as +/usr/share/pacemaker+ depending on your operating system distribution and how you installed the software. If you want to modify just one section of the configuration, you can query and replace just that section to avoid modifying any others. .Safely using an editor to modify only the resources section ====== -------- # cibadmin --query --scope resources > tmp.xml # vi tmp.xml # cibadmin --replace --scope resources --xml-file tmp.xml -------- ====== To quickly delete a part of the configuration, identify the object you wish to delete by XML tag and id. For example, you might search the CIB for all STONITH-related configuration: .Searching for STONITH-related configuration items ====== ---- # cibadmin --query | grep stonith ---- ====== If you wanted to delete the +primitive+ tag with id +child_DoFencing+, you would run: ---- # cibadmin --delete --xml-text '' ---- See the cibadmin man page for more options. [IMPORTANT] ==== Never edit the live +cib.xml+ file directly. Pacemaker will detect such changes and refuse to use the configuration. ==== [[s-crm_shadow]] == Batch Configuration Changes with crm_shadow == indexterm:[Command-line tool,crm_shadow] Often, it is desirable to preview the effects of a series of configuration changes before updating the live configuration all at once. For this purpose, `crm_shadow` creates a "shadow" copy of the configuration and arranges for all the command-line tools to use it. To begin, simply invoke `crm_shadow --create` with a name of your choice, and follow the simple on-screen instructions. Shadow copies are identified with a name to make it possible to have more than one. [WARNING] ==== Read this section and the on-screen instructions carefully; failure to do so could result in destroying the cluster's active configuration! ==== .Creating and displaying the active sandbox ====== ---- # crm_shadow --create test Setting up shadow instance Type Ctrl-D to exit the crm_shadow shell shadow[test]: shadow[test] # crm_shadow --which test ---- ====== From this point on, all cluster commands will automatically use the shadow copy instead of talking to the cluster's active configuration. Once you have finished experimenting, you can either make the changes active via the `--commit` option, or discard them using the `--delete` option. Again, be sure to follow the on-screen instructions carefully! For a full list of `crm_shadow` options and commands, invoke it with the `--help` option. .Use sandbox to make multiple changes all at once, discard them, and verify real configuration is untouched ====== ---- shadow[test] # crm_failcount -r rsc_c001n01 -G scope=status name=fail-count-rsc_c001n01 value=0 shadow[test] # crm_standby --node c001n02 -v on shadow[test] # crm_standby --node c001n02 -G scope=nodes name=standby value=on shadow[test] # cibadmin --erase --force shadow[test] # cibadmin --query shadow[test] # crm_shadow --delete test --force Now type Ctrl-D to exit the crm_shadow shell shadow[test] # exit # crm_shadow --which No active shadow configuration defined # cibadmin -Q ---- ====== See the next section, <>, for how to test your changes before committing them to the live cluster. [[s-crm_simulate]] == Simulate Cluster Activity with crm_simulate == indexterm:[Command-line tool,crm_simulate] The command-line tool `crm_simulate` shows the results of the same logic the cluster itself uses to respond to a particular cluster configuration and status. As always, the man page is the primary documentation, and should be consulted for further details. This section aims for a better conceptual explanation and practical examples. === Replaying cluster decision-making logic === At any given time, one node in a Pacemaker cluster will be elected DC, and that node will run Pacemaker's scheduler to make decisions. Each time decisions need to be made (a "transition"), the DC will have log messages like "Calculated transition ... saving inputs in ..." with a file name. You can grab the named file and replay the cluster logic to see why particular decisions were made. The file contains the live cluster configuration at that moment, so you can also look at it directly to see the value of node attributes, etc., at that time. The simplest usage is (replacing $FILENAME with the actual file name): .Simulate cluster response to a given CIB ==== ---- crm_simulate --simulate --xml-file $FILENAME ---- ==== That will show the cluster state when the process started, the actions that need to be taken ("Transition Summary"), and the resulting cluster state if the actions succeed. Most actions will have a brief description of why they were required. The transition inputs may be compressed. `crm_simulate` can handle these compressed files directly, though if you want to edit the file, you'll need to uncompress it first. You can do the same simulation for the live cluster configuration at the current moment. This is useful mainly when using `crm_shadow` to create a sandbox version of the CIB; the `--live-check` option will use the shadow CIB if one is in effect. .Simulate cluster response to current live CIB or shadow CIB ==== ---- crm_simulate --simulate --live-check ---- ==== === Why decisions were made === To get further insight into the "why", it gets user-unfriendly very quickly. If you add the `--show-scores` option, you will also see all the scores that went into the decision-making. The node with the highest cumulative score for a resource will run it. You can look for +-INFINITY+ scores in particular to see where complete bans came into effect. You can also add `-VVVV` to get more detailed messages about what's happening under the hood. You can add up to two more V's even, but that's usually useful only if you're a masochist or tracing through the source code. === Visualizing the action sequence === Another handy feature is the ability to generate a visual graph of the actions needed, using the `--dot-file` option. This relies on the separate Graphviz footnote:[Graph visualization software. See http://www.graphviz.org/ for details.] project. .Generate a visual graph of cluster actions from a saved CIB ==== ---- crm_simulate --simulate --xml-file $FILENAME --dot-file $FILENAME.dot dot $FILENAME.dot -Tsvg > $FILENAME.svg ---- ==== +$FILENAME.dot+ will contain a GraphViz representation of the cluster's response to your changes, including all actions with their ordering dependencies. +$FILENAME.svg+ will be the same information in a standard graphical format that you can view in your browser or other app of choice. You could, of course, use other `dot` options to generate other formats. How to interpret the graphical output: * Bubbles indicate actions, and arrows indicate ordering dependencies * Resource actions have text of the form pass:[resource]_pass:[action]_pass:[interval] pass:[node] indicating that the specified action will be executed for the specified resource on the specified node, once if interval is 0 or at specified recurring milliseconds interval otherwise * Actions with black text will be sent to the executor (that is, the appropriate agent will be invoked) * Actions with orange text are "pseudo" actions that the cluster uses internally for ordering but require no real activity * Actions with a solid green border are part of the transition (that is, the cluster will attempt to execute them in the given order -- though a transition can be interrupted by action failure or new events) * Dashed arrows indicate dependencies that are not present in the transition graph * Actions with a dashed border will not be executed. If the dashed border is blue, the cluster does not feel the action needs to be executed. If the dashed border is red, the cluster would like to execute the action but cannot. Any actions depending on an action with a dashed border will not be able to execute. * Loops should not happen, and should be reported as a bug if found. .Small Cluster Transition ==== image::images/Policy-Engine-small.png["An example transition graph as represented by Graphviz",width="1161",height="325",align="center"] ==== In the above example, it appears that a new node, *pcmk-2*, has come online and that the cluster is checking to make sure *rsc1*, *rsc2* and *rsc3* are not already running there (indicated by the *rscN_monitor_0* entries). Once it did that, and assuming the resources were not active there, it would have liked to stop *rsc1* and *rsc2* on *pcmk-1* and move them to *pcmk-2*. However, there appears to be some problem and the cluster cannot or is not permitted to perform the stop actions which implies it also cannot perform the start actions. For some reason, the cluster does not want to start *rsc3* anywhere. .Complex Cluster Transition ==== image::images/Policy-Engine-big.png["Complex transition graph that you're not expected to be able to read",width="1455",height="1945",align="center"] ==== === What-if scenarios === You can make changes to the saved or shadow CIB and simulate it again, to see how Pacemaker would react differently. You can edit the XML by hand, use command-line tools such as `cibadmin` with either a shadow CIB or the +CIB_file+ environment variable set to the filename, or use higher-level tool support (see the man pages of the specific tool you're using for how to perform actions on a saved CIB file rather than the live CIB). You can also inject node failures and/or action failures into the simulation; see the `crm_simulate` man page for more details. This capability is useful when using a shadow CIB to edit the configuration. Before committing the changes to the live cluster with `crm_shadow --commit`, you can use `crm_simulate` to see how the cluster will react to the changes. [[s-attrd_updater]] [[s-crm_attribute]] == Manage Node Attributes, Cluster Options and Defaults with crm_attribute and attrd_updater == indexterm:[Command-line tool,attrd_updater] indexterm:[Command-line tool,crm_attribute] `crm_attribute` and `attrd_updater` are confusingly similar tools with subtle differences. `attrd_updater` can query and update node attributes. `crm_attribute` can query and update not only node attributes, but also cluster options, resource defaults, and operation defaults. To understand the differences, it helps to understand the various types of node attribute. .Types of Node Attributes [width="95%",cols="1,1,1,1,1,1",options="header",align="center"] |==== |Type |Recorded in CIB? |Recorded in attribute manager memory? |Survive full cluster restart? |Manageable by crm_attribute? |Manageable by attrd_updater? |permanent |yes |no |yes |yes |no |transient |yes |yes |no |yes |yes |private |no |yes |no |no |yes |==== As you can see from the table above, `crm_attribute` can manage permanent and transient node attributes, while `attrd_updater` can manage transient and private node attributes. The difference between the two tools lies mainly in 'how' they update node attributes: `attrd_updater` always contacts the Pacemaker attribute manager directly, while `crm_attribute` will contact the attribute manager only for transient node attributes, and will instead modify the CIB directly for permanent node attributes (and for transient node attributes when unable to contact the attribute manager). By contacting the attribute manager directly, `attrd_updater` can change an attribute's "dampening" (whether changes are immediately flushed to the CIB or after a specified amount of time, to minimize disk writes for frequent changes), set private node attributes (which are never written to the CIB), and set attributes for nodes that don't yet exist. By modifying the CIB directly, `crm_attribute` can set permanent node attributes (which are only in the CIB and not managed by the attribute manager), and can be used with saved CIB files and shadow CIBs. However a transient node attribute is set, it is synchronized between the CIB and the attribute manager, on all nodes. == Other Commonly Used Tools == Other command-line tools include: indexterm:[Command-line tool,crm_failcount] indexterm:[Command-line tool,crm_node] indexterm:[Command-line tool,crm_report] indexterm:[Command-line tool,crm_standby] indexterm:[Command-line tool,crm_verify] indexterm:[Command-line tool,stonith_admin] * `crm_failcount`: query or delete resource fail counts * `crm_node`: manage cluster nodes * `crm_report`: generate a detailed cluster report for bug submissions * `crm_resource`: manage cluster resources * `crm_standby`: manage standby status of nodes * `crm_verify`: validate a CIB * `stonith_admin`: manage fencing devices See the manual pages for details. diff --git a/tools/crm_mon.c b/tools/crm_mon.c index 43e40d95eb..51366522f8 100644 --- a/tools/crm_mon.c +++ b/tools/crm_mon.c @@ -1,1885 +1,1896 @@ /* * Copyright 2004-2019 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. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* crm_ends_with_ext */ #include #include #include #include #include #include #include #include #include #include #include #include #include "crm_mon.h" #define SUMMARY "Provides a summary of cluster's current state.\n\n" \ "Outputs varying levels of detail in a number of different formats." /* * Definitions indicating which items to print */ static unsigned int show = mon_show_default; /* * Definitions indicating how to output */ static mon_output_format_t output_format = mon_output_unset; /* other globals */ static GMainLoop *mainloop = NULL; static guint timer_id = 0; static mainloop_timer_t *refresh_timer = NULL; static pe_working_set_t *mon_data_set = NULL; static cib_t *cib = NULL; static stonith_t *st = NULL; static xmlNode *current_cib = NULL; static pcmk__common_args_t *args = NULL; static pcmk__output_t *out = NULL; static GOptionContext *context = NULL; /* FIXME allow, detect, and correctly interpret glob pattern or regex? */ const char *print_neg_location_prefix = ""; static time_t last_refresh = 0; crm_trigger_t *refresh_trigger = NULL; static pcmk__supported_format_t formats[] = { #if CURSES_ENABLED CRM_MON_SUPPORTED_FORMAT_CURSES, #endif PCMK__SUPPORTED_FORMAT_HTML, PCMK__SUPPORTED_FORMAT_NONE, PCMK__SUPPORTED_FORMAT_TEXT, CRM_MON_SUPPORTED_FORMAT_XML, { NULL, NULL, NULL } }; /* Define exit codes for monitoring-compatible output * For nagios plugins, the possibilities are * OK=0, WARN=1, CRIT=2, and UNKNOWN=3 */ #define MON_STATUS_WARN CRM_EX_ERROR #define MON_STATUS_CRIT CRM_EX_INVALID_PARAM #define MON_STATUS_UNKNOWN CRM_EX_UNIMPLEMENT_FEATURE #define RECONNECT_MSECS 5000 struct { int reconnect_msec; int fence_history_level; gboolean daemonize; gboolean show_bans; char *pid_file; char *external_agent; char *external_recipient; unsigned int mon_ops; } options = { .reconnect_msec = RECONNECT_MSECS, .fence_history_level = 1, .mon_ops = mon_op_default }; static void clean_up_connections(void); static crm_exit_t clean_up(crm_exit_t exit_code); static void crm_diff_update(const char *event, xmlNode * msg); static gboolean mon_refresh_display(gpointer user_data); static int cib_connect(gboolean full); static void mon_st_callback_event(stonith_t * st, stonith_event_t * e); static void mon_st_callback_display(stonith_t * st, stonith_event_t * e); static void kick_refresh(gboolean data_updated); static gboolean as_cgi_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("html"); output_format = mon_output_cgi; options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean as_html_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { if (args->output_ty != NULL) { free(args->output_ty); } if (args->output_dest != NULL) { free(args->output_dest); } if (optarg != NULL) { args->output_dest = strdup(optarg); } args->output_ty = strdup("html"); output_format = mon_output_html; umask(S_IWGRP | S_IWOTH); return TRUE; } static gboolean as_simple_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("text"); output_format = mon_output_monitor; options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean as_xml_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("xml"); output_format = mon_output_legacy_xml; options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean fence_history_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { int rc = crm_atoi(optarg, "2"); if (rc == -1 || rc > 3) { g_set_error(error, G_OPTION_ERROR, CRM_EX_INVALID_PARAM, "Fence history must be 0-3"); return FALSE; } else { options.fence_history_level = rc; } return TRUE; } static gboolean group_by_node_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_group_by_node; return TRUE; } static gboolean hide_headers_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show &= ~mon_show_headers; return TRUE; } static gboolean inactive_resources_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_inactive_resources; return TRUE; } static gboolean no_curses_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { output_format = mon_output_plain; return TRUE; } static gboolean one_shot_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean print_brief_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_print_brief; return TRUE; } static gboolean print_clone_detail_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_print_clone_detail; return TRUE; } static gboolean print_pending_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_print_pending; return TRUE; } static gboolean print_timing_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_print_timing; show |= mon_show_operations; return TRUE; } static gboolean reconnect_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { int rc = crm_get_msec(optarg); if (rc == -1) { g_set_error(error, G_OPTION_ERROR, CRM_EX_INVALID_PARAM, "Invalid value for -i: %s", optarg); return FALSE; } else { options.reconnect_msec = crm_get_msec(optarg); } return TRUE; } static gboolean show_attributes_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_attributes; return TRUE; } static gboolean show_bans_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_bans; if (optarg != NULL) { print_neg_location_prefix = optarg; } return TRUE; } static gboolean show_failcounts_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_failcounts; return TRUE; } static gboolean show_operations_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_operations; return TRUE; } static gboolean show_tickets_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { show |= mon_show_tickets; return TRUE; } static gboolean use_cib_file_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { setenv("CIB_file", optarg, 1); options.mon_ops |= mon_op_one_shot; return TRUE; } static gboolean watch_fencing_cb(const gchar *option_name, const gchar *optarg, gpointer data, GError **error) { options.mon_ops |= mon_op_watch_fencing; return TRUE; } #define INDENT " " /* *INDENT-OFF* */ static GOptionEntry addl_entries[] = { { "interval", 'i', 0, G_OPTION_ARG_CALLBACK, reconnect_cb, "Update frequency in seconds (default is 5)", "SECONDS" }, { "one-shot", '1', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, one_shot_cb, "Display the cluster status once on the console and exit", NULL }, { "daemonize", 'd', 0, G_OPTION_ARG_NONE, &options.daemonize, "Run in the background as a daemon.\n" INDENT "Requires at least one of --output-to and --external-agent.", NULL }, { "pid-file", 'p', 0, G_OPTION_ARG_FILENAME, &options.pid_file, "(Advanced) Daemon pid file location", "FILE" }, { "external-agent", 'E', 0, G_OPTION_ARG_FILENAME, &options.external_agent, "A program to run when resource operations take place", "FILE" }, { "external-recipient", 'e', 0, G_OPTION_ARG_STRING, &options.external_recipient, "A recipient for your program (assuming you want the program to send something to someone).", "RCPT" }, { "xml-file", 'x', G_OPTION_FLAG_HIDDEN, G_OPTION_ARG_CALLBACK, use_cib_file_cb, NULL, NULL }, { NULL } }; static GOptionEntry display_entries[] = { { "group-by-node", 'n', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, group_by_node_cb, "Group resources by node", NULL }, { "inactive", 'r', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, inactive_resources_cb, "Display inactive resources", NULL }, { "failcounts", 'f', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, show_failcounts_cb, "Display resource fail counts", NULL }, { "operations", 'o', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, show_operations_cb, "Display resource operation history", NULL }, { "timing-details", 't', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, print_timing_cb, "Display resource operation history with timing details", NULL }, { "tickets", 'c', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, show_tickets_cb, "Display cluster tickets", NULL }, { "watch-fencing", 'W', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, watch_fencing_cb, "Listen for fencing events. For use with --external-agent", NULL }, { "fence-history", 'm', G_OPTION_FLAG_OPTIONAL_ARG, G_OPTION_ARG_CALLBACK, fence_history_cb, "Show fence history:\n" INDENT "0=off, 1=failures and pending (default without option),\n" INDENT "2=add successes (default without value for option),\n" INDENT "3=show full history without reduction to most recent of each flavor", "LEVEL" }, { "neg-locations", 'L', G_OPTION_FLAG_OPTIONAL_ARG, G_OPTION_ARG_CALLBACK, show_bans_cb, "Display negative location constraints [optionally filtered by id prefix]", NULL }, { "show-node-attributes", 'A', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, show_attributes_cb, "Display node attributes", NULL }, { "hide-headers", 'D', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, hide_headers_cb, "Hide all headers", NULL }, { "show-detail", 'R', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, print_clone_detail_cb, "Show more details (node IDs, individual clone instances)", NULL }, { "brief", 'b', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, print_brief_cb, "Brief output", NULL }, { "pending", 'j', G_OPTION_FLAG_HIDDEN|G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, print_pending_cb, "Display pending state if 'record-pending' is enabled", NULL }, { "simple-status", 's', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, as_simple_cb, "Display the cluster status once as a simple one line output (suitable for nagios)", NULL }, { NULL } }; static GOptionEntry deprecated_entries[] = { { "as-html", 'h', G_OPTION_FLAG_FILENAME, G_OPTION_ARG_CALLBACK, as_html_cb, "Write cluster status to the named HTML file.\n" INDENT "Use --output-as=html --output-to=FILE instead.", "FILE" }, { "as-xml", 'X', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, as_xml_cb, "Write cluster status as XML to stdout. This will enable one-shot mode.\n" INDENT "Use --output-as=xml instead.", NULL }, { "disable-ncurses", 'N', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, no_curses_cb, "Disable the use of ncurses.\n" INDENT "Use --output-as=text instead.", NULL }, { "web-cgi", 'w', G_OPTION_FLAG_NO_ARG, G_OPTION_ARG_CALLBACK, as_cgi_cb, "Web mode with output suitable for CGI (preselected when run as *.cgi).\n" INDENT "Use --output-as=html --output-cgi instead.", NULL }, { NULL } }; /* *INDENT-ON* */ static gboolean mon_timer_popped(gpointer data) { int rc = pcmk_ok; #if CURSES_ENABLED if (output_format == mon_output_console) { clear(); refresh(); } #endif if (timer_id > 0) { g_source_remove(timer_id); timer_id = 0; } print_as(output_format, "Reconnecting...\n"); rc = cib_connect(TRUE); if (rc != pcmk_ok) { timer_id = g_timeout_add(options.reconnect_msec, mon_timer_popped, NULL); } return FALSE; } static void mon_cib_connection_destroy(gpointer user_data) { print_as(output_format, "Connection to the cluster-daemons terminated\n"); if (refresh_timer != NULL) { /* we'll trigger a refresh after reconnect */ mainloop_timer_stop(refresh_timer); } if (timer_id) { /* we'll trigger a new reconnect-timeout at the end */ g_source_remove(timer_id); timer_id = 0; } if (st) { /* the client API won't properly reconnect notifications * if they are still in the table - so remove them */ st->cmds->remove_notification(st, T_STONITH_NOTIFY_DISCONNECT); st->cmds->remove_notification(st, T_STONITH_NOTIFY_FENCE); st->cmds->remove_notification(st, T_STONITH_NOTIFY_HISTORY); if (st->state != stonith_disconnected) { st->cmds->disconnect(st); } } if (cib) { cib->cmds->signoff(cib); timer_id = g_timeout_add(options.reconnect_msec, mon_timer_popped, NULL); } return; } /* * Mainloop signal handler. */ static void mon_shutdown(int nsig) { clean_up(CRM_EX_OK); } #if CURSES_ENABLED static sighandler_t ncurses_winch_handler; static void mon_winresize(int nsig) { static int not_done; int lines = 0, cols = 0; if (!not_done++) { if (ncurses_winch_handler) /* the original ncurses WINCH signal handler does the * magic of retrieving the new window size; * otherwise, we'd have to use ioctl or tgetent */ (*ncurses_winch_handler) (SIGWINCH); getmaxyx(stdscr, lines, cols); resizeterm(lines, cols); mainloop_set_trigger(refresh_trigger); } not_done--; } #endif static int cib_connect(gboolean full) { int rc = pcmk_ok; static gboolean need_pass = TRUE; CRM_CHECK(cib != NULL, return -EINVAL); if (getenv("CIB_passwd") != NULL) { need_pass = FALSE; } if (is_set(options.mon_ops, mon_op_fence_connect) && st == NULL) { st = stonith_api_new(); } if (is_set(options.mon_ops, mon_op_fence_connect) && st != NULL && st->state == stonith_disconnected) { rc = st->cmds->connect(st, crm_system_name, NULL); if (rc == pcmk_ok) { crm_trace("Setting up stonith callbacks"); if (is_set(options.mon_ops, mon_op_watch_fencing)) { st->cmds->register_notification(st, T_STONITH_NOTIFY_DISCONNECT, mon_st_callback_event); st->cmds->register_notification(st, T_STONITH_NOTIFY_FENCE, mon_st_callback_event); } else { st->cmds->register_notification(st, T_STONITH_NOTIFY_DISCONNECT, mon_st_callback_display); st->cmds->register_notification(st, T_STONITH_NOTIFY_HISTORY, mon_st_callback_display); } } } if (cib->state != cib_connected_query && cib->state != cib_connected_command) { crm_trace("Connecting to the CIB"); if ((output_format == mon_output_console) && need_pass && (cib->variant == cib_remote)) { need_pass = FALSE; print_as(output_format, "Password:"); } rc = cib->cmds->signon(cib, crm_system_name, cib_query); if (rc != pcmk_ok) { return rc; } rc = cib->cmds->query(cib, NULL, ¤t_cib, cib_scope_local | cib_sync_call); if (rc == pcmk_ok) { mon_refresh_display(&output_format); } if (rc == pcmk_ok && full) { if (rc == pcmk_ok) { rc = cib->cmds->set_connection_dnotify(cib, mon_cib_connection_destroy); if (rc == -EPROTONOSUPPORT) { print_as (output_format, "Notification setup not supported, won't be able to reconnect after failure"); if (output_format == mon_output_console) { sleep(2); } rc = pcmk_ok; } } if (rc == pcmk_ok) { cib->cmds->del_notify_callback(cib, T_CIB_DIFF_NOTIFY, crm_diff_update); rc = cib->cmds->add_notify_callback(cib, T_CIB_DIFF_NOTIFY, crm_diff_update); } if (rc != pcmk_ok) { print_as(output_format, "Notification setup failed, could not monitor CIB actions"); if (output_format == mon_output_console) { sleep(2); } clean_up_connections(); } } } return rc; } #if CURSES_ENABLED static const char * get_option_desc(char c) { const char *desc = "No help available"; for (GOptionEntry *entry = display_entries; entry != NULL; entry++) { if (entry->short_name == c) { desc = entry->description; break; } } return desc; } #define print_option_help(output_format, option, condition) \ - print_as(output_format, "%c %c: \t%s\n", ((condition)? '*': ' '), option, get_option_desc(option)); + out->info(out, "%c %c: \t%s", ((condition)? '*': ' '), option, get_option_desc(option)); static gboolean detect_user_input(GIOChannel *channel, GIOCondition condition, gpointer user_data) { int c; gboolean config_mode = FALSE; while (1) { /* Get user input */ c = getchar(); switch (c) { case 'm': if (!options.fence_history_level) { options.mon_ops |= mon_op_fence_history; options.mon_ops |= mon_op_fence_connect; if (st == NULL) { mon_cib_connection_destroy(NULL); } } show ^= mon_show_fence_history; break; case 'c': show ^= mon_show_tickets; break; case 'f': show ^= mon_show_failcounts; break; case 'n': options.mon_ops ^= mon_op_group_by_node; break; case 'o': show ^= mon_show_operations; if ((show & mon_show_operations) == 0) { options.mon_ops &= ~mon_op_print_timing; } break; case 'r': options.mon_ops ^= mon_op_inactive_resources; break; case 'R': options.mon_ops ^= mon_op_print_clone_detail; break; case 't': options.mon_ops ^= mon_op_print_timing; if (is_set(options.mon_ops, mon_op_print_timing)) { show |= mon_show_operations; } break; case 'A': show ^= mon_show_attributes; break; case 'L': show ^= mon_show_bans; break; case 'D': /* If any header is shown, clear them all, otherwise set them all */ if (show & mon_show_headers) { show &= ~mon_show_headers; } else { show |= mon_show_headers; } break; case 'b': options.mon_ops ^= mon_op_print_brief; break; case 'j': options.mon_ops ^= mon_op_print_pending; break; case '?': config_mode = TRUE; break; default: goto refresh; } if (!config_mode) goto refresh; blank_screen(); - print_as(output_format, "Display option change mode\n"); - print_as(output_format, "\n"); - print_option_help(output_format, 'c', show & mon_show_tickets); - print_option_help(output_format, 'f', show & mon_show_failcounts); - print_option_help(output_format, 'n', is_set(options.mon_ops, mon_op_group_by_node)); - print_option_help(output_format, 'o', show & mon_show_operations); - print_option_help(output_format, 'r', is_set(options.mon_ops, mon_op_inactive_resources)); - print_option_help(output_format, 't', is_set(options.mon_ops, mon_op_print_timing)); - print_option_help(output_format, 'A', show & mon_show_attributes); - print_option_help(output_format, 'L', show & mon_show_bans); - print_option_help(output_format, 'D', (show & mon_show_headers) == 0); - print_option_help(output_format, 'R', is_set(options.mon_ops, mon_op_print_clone_detail)); - print_option_help(output_format, 'b', is_set(options.mon_ops, mon_op_print_brief)); - print_option_help(output_format, 'j', is_set(options.mon_ops, mon_op_print_pending)); - print_option_help(output_format, 'm', (show & mon_show_fence_history)); - print_as(output_format, "\n"); - print_as(output_format, "Toggle fields via field letter, type any other key to return"); + out->info(out, "%s", "Display option change mode\n"); + print_option_help(out, 'c', show & mon_show_tickets); + print_option_help(out, 'f', show & mon_show_failcounts); + print_option_help(out, 'n', is_set(options.mon_ops, mon_op_group_by_node)); + print_option_help(out, 'o', show & mon_show_operations); + print_option_help(out, 'r', is_set(options.mon_ops, mon_op_inactive_resources)); + print_option_help(out, 't', is_set(options.mon_ops, mon_op_print_timing)); + print_option_help(out, 'A', show & mon_show_attributes); + print_option_help(out, 'L', show & mon_show_bans); + print_option_help(out, 'D', (show & mon_show_headers) == 0); + print_option_help(out, 'R', is_set(options.mon_ops, mon_op_print_clone_detail)); + print_option_help(out, 'b', is_set(options.mon_ops, mon_op_print_brief)); + print_option_help(out, 'j', is_set(options.mon_ops, mon_op_print_pending)); + print_option_help(out, 'm', (show & mon_show_fence_history)); + out->info(out, "%s", "\nToggle fields via field letter, type any other key to return"); } refresh: mon_refresh_display(NULL); return TRUE; } #endif // Basically crm_signal_handler(SIGCHLD, SIG_IGN) plus the SA_NOCLDWAIT flag static void avoid_zombies() { struct sigaction sa; memset(&sa, 0, sizeof(struct sigaction)); if (sigemptyset(&sa.sa_mask) < 0) { crm_warn("Cannot avoid zombies: %s", pcmk_strerror(errno)); return; } sa.sa_handler = SIG_IGN; sa.sa_flags = SA_RESTART|SA_NOCLDWAIT; if (sigaction(SIGCHLD, &sa, NULL) < 0) { crm_warn("Cannot avoid zombies: %s", pcmk_strerror(errno)); } } static GOptionContext * build_arg_context(pcmk__common_args_t *args) { GOptionContext *context = NULL; GOptionEntry extra_prog_entries[] = { { "quiet", 'Q', 0, G_OPTION_ARG_NONE, &(args->quiet), "Be less descriptive in output.", NULL }, { NULL } }; const char *examples = "Examples:\n\n" "Display the cluster status on the console with updates as they occur:\n\n" "\tcrm_mon\n\n" "Display the cluster status on the console just once then exit:\n\n" "\tcrm_mon -1\n\n" "Display your cluster status, group resources by node, and include inactive resources in the list:\n\n" "\tcrm_mon --group-by-node --inactive\n\n" "Start crm_mon as a background daemon and have it write the cluster status to an HTML file:\n\n" "\tcrm_mon --daemonize --output-as html --output-to /path/to/docroot/filename.html\n\n" "Start crm_mon and export the current cluster status as XML to stdout, then exit:\n\n" "\tcrm_mon --output-as xml\n"; context = pcmk__build_arg_context(args, "console (default), html, text, xml"); pcmk__add_main_args(context, extra_prog_entries); g_option_context_set_description(context, examples); pcmk__add_arg_group(context, "display", "Display Options:", "Show display options", display_entries); pcmk__add_arg_group(context, "additional", "Additional Options:", "Show additional options", addl_entries); pcmk__add_arg_group(context, "deprecated", "Deprecated Options:", "Show deprecated options", deprecated_entries); return context; } /* If certain format options were specified, we want to set some extra * options. We can just process these like they were given on the * command line. */ static void add_output_args() { GError *error = NULL; if (output_format == mon_output_plain) { if (!pcmk__force_args(context, &error, "%s --output-fancy", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } else if (output_format == mon_output_html) { if (!pcmk__force_args(context, &error, "%s --output-title \"Cluster Status\"", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } else if (output_format == mon_output_cgi) { if (!pcmk__force_args(context, &error, "%s --output-cgi --output-title \"Cluster Status\"", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } else if (output_format == mon_output_xml) { if (!pcmk__force_args(context, &error, "%s --output-simple-list", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } else if (output_format == mon_output_legacy_xml) { output_format = mon_output_xml; if (!pcmk__force_args(context, &error, "%s --output-legacy-xml", g_get_prgname())) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } } /* Which output format to use could come from two places: The --as-xml * style arguments we gave in deprecated_entries above, or the formatted output * arguments added by pcmk__register_formats. If the latter were used, * output_format will be mon_output_unset. * * Call the callbacks as if those older style arguments were provided so * the various things they do get done. */ static void reconcile_output_format(pcmk__common_args_t *args) { gboolean retval = TRUE; GError *error = NULL; if (output_format != mon_output_unset) { return; } if (safe_str_eq(args->output_ty, "html")) { char *dest = NULL; if (args->output_dest != NULL) { dest = strdup(args->output_dest); } retval = as_html_cb("h", dest, NULL, &error); free(dest); } else if (safe_str_eq(args->output_ty, "text")) { retval = no_curses_cb("N", NULL, NULL, &error); } else if (safe_str_eq(args->output_ty, "xml")) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("xml"); output_format = mon_output_xml; options.mon_ops |= mon_op_one_shot; } else if (is_set(options.mon_ops, mon_op_one_shot)) { if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("text"); output_format = mon_output_plain; } else { /* Neither old nor new arguments were given, so set the default. */ if (args->output_ty != NULL) { free(args->output_ty); } args->output_ty = strdup("console"); output_format = mon_output_console; } if (!retval) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); clean_up(CRM_EX_USAGE); } } int main(int argc, char **argv) { int rc = pcmk_ok; char **processed_args = NULL; GError *error = NULL; args = pcmk__new_common_args(SUMMARY); context = build_arg_context(args); pcmk__register_formats(context, formats); options.pid_file = strdup("/tmp/ClusterMon.pid"); crm_log_cli_init("crm_mon"); // Avoid needing to wait for subprocesses forked for -E/--external-agent avoid_zombies(); if (crm_ends_with_ext(argv[0], ".cgi") == TRUE) { output_format = mon_output_cgi; options.mon_ops |= mon_op_one_shot; } processed_args = pcmk__cmdline_preproc(argc, argv, "ehimpxEL"); if (!g_option_context_parse_strv(context, &processed_args, &error)) { fprintf(stderr, "%s: %s\n", g_get_prgname(), error->message); return clean_up(CRM_EX_USAGE); } for (int i = 0; i < args->verbosity; i++) { crm_bump_log_level(argc, argv); } if (!args->version) { if (args->quiet) { show &= ~mon_show_times; } if (is_set(options.mon_ops, mon_op_watch_fencing)) { options.mon_ops |= mon_op_fence_connect; /* don't moan as fence_history_level == 1 is default */ options.fence_history_level = 0; } /* create the cib-object early to be able to do further * decisions based on the cib-source */ cib = cib_new(); if (cib == NULL) { rc = -EINVAL; } else { switch (cib->variant) { case cib_native: /* cib & fencing - everything available */ break; case cib_file: /* Don't try to connect to fencing as we * either don't have a running cluster or * the fencing-information would possibly * not match the cib data from a file. * As we don't expect cib-updates coming * in enforce one-shot. */ options.fence_history_level = 0; options.mon_ops |= mon_op_one_shot; break; case cib_remote: /* updates coming in but no fencing */ options.fence_history_level = 0; break; case cib_undefined: case cib_database: default: /* something is odd */ rc = -EINVAL; break; } } switch (options.fence_history_level) { case 3: options.mon_ops |= mon_op_fence_full_history; /* fall through to next lower level */ case 2: show |= mon_show_fence_history; /* fall through to next lower level */ case 1: options.mon_ops |= mon_op_fence_history; options.mon_ops |= mon_op_fence_connect; break; default: break; } if (is_set(options.mon_ops, mon_op_one_shot)) { if (output_format == mon_output_console) { output_format = mon_output_plain; } } else if (options.daemonize) { if ((output_format == mon_output_console) || (output_format == mon_output_plain)) { output_format = mon_output_none; } crm_enable_stderr(FALSE); if ((args->output_dest == NULL || safe_str_eq(args->output_dest, "-")) && !options.external_agent) { printf("--daemonize requires at least one of --output-to and --external-agent\n"); return clean_up(CRM_EX_USAGE); } if (cib) { /* to be on the safe side don't have cib-object around * when we are forking */ cib_delete(cib); cib = NULL; crm_make_daemon(crm_system_name, TRUE, options.pid_file); cib = cib_new(); if (cib == NULL) { rc = -EINVAL; } /* otherwise assume we've got the same cib-object we've just destroyed * in our parent */ } } else if (output_format == mon_output_console) { #if CURSES_ENABLED crm_enable_stderr(FALSE); #else options.mon_ops |= mon_op_one_shot; output_format = mon_output_plain; printf("Defaulting to one-shot mode\n"); printf("You need to have curses available at compile time to enable console mode\n"); #endif } } if (rc != pcmk_ok) { // Shouldn't really be possible fprintf(stderr, "Invalid CIB source\n"); return clean_up(CRM_EX_ERROR); } reconcile_output_format(args); add_output_args(); /* Create the output format - output_format must not be changed after this point. */ if (args->version && output_format == mon_output_console) { /* Use the text output format here if we are in curses mode but were given * --version. Displaying version information uses printf, and then we * immediately exit. We don't want to initialize curses for that. */ rc = pcmk__output_new(&out, "text", args->output_dest, argv); } else { rc = pcmk__output_new(&out, args->output_ty, args->output_dest, argv); } if (rc != 0) { fprintf(stderr, "Error creating output format %s: %s\n", args->output_ty, pcmk_strerror(rc)); return clean_up(CRM_EX_ERROR); } crm_mon_register_messages(out); pe__register_messages(out); stonith__register_messages(out); if (args->version) { out->version(out, false); return clean_up(CRM_EX_OK); } /* Extra sanity checks when in CGI mode */ if (output_format == mon_output_cgi) { if (cib && cib->variant == cib_file) { fprintf(stderr, "CGI mode used with CIB file\n"); return clean_up(CRM_EX_USAGE); } else if (options.external_agent != NULL) { fprintf(stderr, "CGI mode cannot be used with --external-agent\n"); return clean_up(CRM_EX_USAGE); } else if (options.daemonize == TRUE) { fprintf(stderr, "CGI mode cannot be used with -d\n"); return clean_up(CRM_EX_USAGE); } } /* XML output always prints everything */ if (output_format == mon_output_xml || output_format == mon_output_legacy_xml) { show = mon_show_all; options.mon_ops |= mon_op_print_timing; } crm_info("Starting %s", crm_system_name); if (cib) { do { if (is_not_set(options.mon_ops, mon_op_one_shot)) { print_as(output_format ,"Waiting until cluster is available on this node ...\n"); } rc = cib_connect(is_not_set(options.mon_ops, mon_op_one_shot)); if (is_set(options.mon_ops, mon_op_one_shot)) { break; } else if (rc != pcmk_ok) { sleep(options.reconnect_msec / 1000); #if CURSES_ENABLED if (output_format == mon_output_console) { clear(); refresh(); } #endif } else { if (output_format == mon_output_html && out->dest != stdout) { printf("Writing html to %s ...\n", args->output_dest); } } } while (rc == -ENOTCONN); } if (rc != pcmk_ok) { if (output_format == mon_output_monitor) { printf("CLUSTER CRIT: Connection to cluster failed: %s\n", pcmk_strerror(rc)); return clean_up(MON_STATUS_CRIT); } else { if (rc == -ENOTCONN) { - print_as(output_format ,"\nError: cluster is not available on this node\n"); + out->err(out, "%s", "\nError: cluster is not available on this node"); } else { - print_as(output_format ,"\nConnection to cluster failed: %s\n", - pcmk_strerror(rc)); + out->err(out, "\nConnection to cluster failed: %s", pcmk_strerror(rc)); } } if (output_format == mon_output_console) { sleep(2); } return clean_up(crm_errno2exit(rc)); } if (is_set(options.mon_ops, mon_op_one_shot)) { return clean_up(CRM_EX_OK); } mainloop = g_main_loop_new(NULL, FALSE); mainloop_add_signal(SIGTERM, mon_shutdown); mainloop_add_signal(SIGINT, mon_shutdown); #if CURSES_ENABLED if (output_format == mon_output_console) { ncurses_winch_handler = crm_signal_handler(SIGWINCH, mon_winresize); if (ncurses_winch_handler == SIG_DFL || ncurses_winch_handler == SIG_IGN || ncurses_winch_handler == SIG_ERR) ncurses_winch_handler = NULL; g_io_add_watch(g_io_channel_unix_new(STDIN_FILENO), G_IO_IN, detect_user_input, NULL); } #endif refresh_trigger = mainloop_add_trigger(G_PRIORITY_LOW, mon_refresh_display, NULL); g_main_loop_run(mainloop); g_main_loop_unref(mainloop); crm_info("Exiting %s", crm_system_name); return clean_up(CRM_EX_OK); } -#define mon_warn(output_format, mon_ops, fmt...) do { \ - if (is_not_set(mon_ops, mon_op_has_warnings)) { \ - print_as(output_format, "CLUSTER WARN:"); \ - } else { \ - print_as(output_format, ","); \ - } \ - print_as(output_format, fmt); \ - mon_ops |= mon_op_has_warnings; \ - } while(0) - /*! * \internal * \brief Print one-line status suitable for use with monitoring software * * \param[in] data_set Working set of CIB state * \param[in] history List of stonith actions * * \note This function's output (and the return code when the program exits) * should conform to https://www.monitoring-plugins.org/doc/guidelines.html */ static void -print_simple_status(pe_working_set_t * data_set, stonith_history_t *history, - unsigned int mon_ops, mon_output_format_t output_format) +print_simple_status(pcmk__output_t *out, pe_working_set_t * data_set, + stonith_history_t *history, unsigned int mon_ops) { GListPtr gIter = NULL; int nodes_online = 0; int nodes_standby = 0; int nodes_maintenance = 0; + char *offline_nodes = NULL; + gboolean no_dc = FALSE; + gboolean offline = FALSE; if (data_set->dc_node == NULL) { - mon_warn(output_format, mon_ops, " No DC"); + mon_ops |= mon_op_has_warnings; + no_dc = TRUE; } for (gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) { node_t *node = (node_t *) gIter->data; if (node->details->standby && node->details->online) { nodes_standby++; } else if (node->details->maintenance && node->details->online) { nodes_maintenance++; } else if (node->details->online) { nodes_online++; } else { - mon_warn(output_format, mon_ops, " offline node: %s", node->details->uname); + char *s = crm_strdup_printf("offline node: %s", node->details->uname); + offline_nodes = add_list_element(offline_nodes, s); + free(s); + mon_ops |= mon_op_has_warnings; + offline = TRUE; } } - if (is_not_set(mon_ops, mon_op_has_warnings)) { + if (is_set(mon_ops, mon_op_has_warnings)) { + out->info(out, "CLUSTER WARN:%s%s%s", + no_dc ? " No DC" : "", + no_dc && offline ? "," : "", + offline ? offline_nodes : ""); + free(offline_nodes); + } else { int nresources = count_resources(data_set, NULL); + char *nodes_standby_s = NULL; + char *nodes_maint_s = NULL; - printf("CLUSTER OK: %d node%s online", nodes_online, s_if_plural(nodes_online)); if (nodes_standby > 0) { - printf(", %d standby node%s", nodes_standby, s_if_plural(nodes_standby)); + nodes_standby_s = crm_strdup_printf(", %d standby node%s", nodes_standby, + s_if_plural(nodes_standby)); } + if (nodes_maintenance > 0) { - printf(", %d maintenance node%s", nodes_maintenance, s_if_plural(nodes_maintenance)); + nodes_maint_s = crm_strdup_printf(", %d maintenance node%s", nresources, + s_if_plural(nresources)); } - printf(", %d resource%s configured", nresources, s_if_plural(nresources)); - } - printf("\n"); + out->info(out, "CLUSTER OK: %dnode%s online%s%s, %d resource%s configured", + nodes_online, s_if_plural(nodes_online), + nodes_standby_s != NULL ? nodes_standby_s : "", + nodes_maint_s != NULL ? nodes_maint_s : "", + nresources, s_if_plural(nresources)); + + free(nodes_standby_s); + free(nodes_maint_s); + } } /*! * \internal * \brief Reduce the stonith-history * for successful actions we keep the last of every action-type & target * for failed actions we record as well who had failed * for actions in progress we keep full track * * \param[in] history List of stonith actions * */ static stonith_history_t * reduce_stonith_history(stonith_history_t *history) { stonith_history_t *new = history, *hp, *np; if (new) { hp = new->next; new->next = NULL; while (hp) { stonith_history_t *hp_next = hp->next; hp->next = NULL; for (np = new; ; np = np->next) { if ((hp->state == st_done) || (hp->state == st_failed)) { /* action not in progress */ if (safe_str_eq(hp->target, np->target) && safe_str_eq(hp->action, np->action) && (hp->state == np->state) && ((hp->state == st_done) || safe_str_eq(hp->delegate, np->delegate))) { /* purge older hp */ stonith_history_free(hp); break; } } if (!np->next) { np->next = hp; break; } } hp = hp_next; } } return new; } static int send_custom_trap(const char *node, const char *rsc, const char *task, int target_rc, int rc, int status, const char *desc) { pid_t pid; /*setenv needs chars, these are ints */ char *rc_s = crm_itoa(rc); char *status_s = crm_itoa(status); char *target_rc_s = crm_itoa(target_rc); crm_debug("Sending external notification to '%s' via '%s'", options.external_recipient, options.external_agent); if(rsc) { setenv("CRM_notify_rsc", rsc, 1); } if (options.external_recipient) { setenv("CRM_notify_recipient", options.external_recipient, 1); } setenv("CRM_notify_node", node, 1); setenv("CRM_notify_task", task, 1); setenv("CRM_notify_desc", desc, 1); setenv("CRM_notify_rc", rc_s, 1); setenv("CRM_notify_target_rc", target_rc_s, 1); setenv("CRM_notify_status", status_s, 1); pid = fork(); if (pid == -1) { crm_perror(LOG_ERR, "notification fork() failed."); } if (pid == 0) { /* crm_debug("notification: I am the child. Executing the nofitication program."); */ execl(options.external_agent, options.external_agent, NULL); exit(CRM_EX_ERROR); } crm_trace("Finished running custom notification program '%s'.", options.external_agent); free(target_rc_s); free(status_s); free(rc_s); return 0; } static void handle_rsc_op(xmlNode * xml, const char *node_id) { int rc = -1; int status = -1; int target_rc = -1; gboolean notify = TRUE; char *rsc = NULL; char *task = NULL; const char *desc = NULL; const char *magic = NULL; const char *id = NULL; const char *node = NULL; xmlNode *n = xml; xmlNode * rsc_op = xml; if(strcmp((const char*)xml->name, XML_LRM_TAG_RSC_OP) != 0) { xmlNode *cIter; for(cIter = xml->children; cIter; cIter = cIter->next) { handle_rsc_op(cIter, node_id); } return; } id = crm_element_value(rsc_op, XML_LRM_ATTR_TASK_KEY); if (id == NULL) { /* Compatibility with <= 1.1.5 */ id = ID(rsc_op); } magic = crm_element_value(rsc_op, XML_ATTR_TRANSITION_MAGIC); if (magic == NULL) { /* non-change */ return; } if (!decode_transition_magic(magic, NULL, NULL, NULL, &status, &rc, &target_rc)) { crm_err("Invalid event %s detected for %s", magic, id); return; } if (parse_op_key(id, &rsc, &task, NULL) == FALSE) { crm_err("Invalid event detected for %s", id); goto bail; } node = crm_element_value(rsc_op, XML_LRM_ATTR_TARGET); while (n != NULL && safe_str_neq(XML_CIB_TAG_STATE, TYPE(n))) { n = n->parent; } if(node == NULL && n) { node = crm_element_value(n, XML_ATTR_UNAME); } if (node == NULL && n) { node = ID(n); } if (node == NULL) { node = node_id; } if (node == NULL) { crm_err("No node detected for event %s (%s)", magic, id); goto bail; } /* look up where we expected it to be? */ desc = pcmk_strerror(pcmk_ok); if (status == PCMK_LRM_OP_DONE && target_rc == rc) { crm_notice("%s of %s on %s completed: %s", task, rsc, node, desc); if (rc == PCMK_OCF_NOT_RUNNING) { notify = FALSE; } } else if (status == PCMK_LRM_OP_DONE) { desc = services_ocf_exitcode_str(rc); crm_warn("%s of %s on %s failed: %s", task, rsc, node, desc); } else { desc = services_lrm_status_str(status); crm_warn("%s of %s on %s failed: %s", task, rsc, node, desc); } if (notify && options.external_agent) { send_custom_trap(node, rsc, task, target_rc, rc, status, desc); } bail: free(rsc); free(task); } static gboolean mon_trigger_refresh(gpointer user_data) { mainloop_set_trigger(refresh_trigger); return FALSE; } #define NODE_PATT "/lrm[@id=" static char * get_node_from_xpath(const char *xpath) { char *nodeid = NULL; char *tmp = strstr(xpath, NODE_PATT); if(tmp) { tmp += strlen(NODE_PATT); tmp += 1; nodeid = strdup(tmp); tmp = strstr(nodeid, "\'"); CRM_ASSERT(tmp); tmp[0] = 0; } return nodeid; } static void crm_diff_update_v2(const char *event, xmlNode * msg) { xmlNode *change = NULL; xmlNode *diff = get_message_xml(msg, F_CIB_UPDATE_RESULT); for (change = __xml_first_child(diff); change != NULL; change = __xml_next(change)) { const char *name = NULL; const char *op = crm_element_value(change, XML_DIFF_OP); const char *xpath = crm_element_value(change, XML_DIFF_PATH); xmlNode *match = NULL; const char *node = NULL; if(op == NULL) { continue; } else if(strcmp(op, "create") == 0) { match = change->children; } else if(strcmp(op, "move") == 0) { continue; } else if(strcmp(op, "delete") == 0) { continue; } else if(strcmp(op, "modify") == 0) { match = first_named_child(change, XML_DIFF_RESULT); if(match) { match = match->children; } } if(match) { name = (const char *)match->name; } crm_trace("Handling %s operation for %s %p, %s", op, xpath, match, name); if(xpath == NULL) { /* Version field, ignore */ } else if(name == NULL) { crm_debug("No result for %s operation to %s", op, xpath); CRM_ASSERT(strcmp(op, "delete") == 0 || strcmp(op, "move") == 0); } else if(strcmp(name, XML_TAG_CIB) == 0) { xmlNode *state = NULL; xmlNode *status = first_named_child(match, XML_CIB_TAG_STATUS); for (state = __xml_first_child_element(status); state != NULL; state = __xml_next_element(state)) { node = crm_element_value(state, XML_ATTR_UNAME); if (node == NULL) { node = ID(state); } handle_rsc_op(state, node); } } else if(strcmp(name, XML_CIB_TAG_STATUS) == 0) { xmlNode *state = NULL; for (state = __xml_first_child_element(match); state != NULL; state = __xml_next_element(state)) { node = crm_element_value(state, XML_ATTR_UNAME); if (node == NULL) { node = ID(state); } handle_rsc_op(state, node); } } else if(strcmp(name, XML_CIB_TAG_STATE) == 0) { node = crm_element_value(match, XML_ATTR_UNAME); if (node == NULL) { node = ID(match); } handle_rsc_op(match, node); } else if(strcmp(name, XML_CIB_TAG_LRM) == 0) { node = ID(match); handle_rsc_op(match, node); } else if(strcmp(name, XML_LRM_TAG_RESOURCES) == 0) { char *local_node = get_node_from_xpath(xpath); handle_rsc_op(match, local_node); free(local_node); } else if(strcmp(name, XML_LRM_TAG_RESOURCE) == 0) { char *local_node = get_node_from_xpath(xpath); handle_rsc_op(match, local_node); free(local_node); } else if(strcmp(name, XML_LRM_TAG_RSC_OP) == 0) { char *local_node = get_node_from_xpath(xpath); handle_rsc_op(match, local_node); free(local_node); } else { crm_trace("Ignoring %s operation for %s %p, %s", op, xpath, match, name); } } } static void crm_diff_update_v1(const char *event, xmlNode * msg) { /* Process operation updates */ xmlXPathObject *xpathObj = xpath_search(msg, "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_ADDED "//" XML_LRM_TAG_RSC_OP); int lpc = 0, max = numXpathResults(xpathObj); for (lpc = 0; lpc < max; lpc++) { xmlNode *rsc_op = getXpathResult(xpathObj, lpc); handle_rsc_op(rsc_op, NULL); } freeXpathObject(xpathObj); } static void crm_diff_update(const char *event, xmlNode * msg) { int rc = -1; static bool stale = FALSE; gboolean cib_updated = FALSE; xmlNode *diff = get_message_xml(msg, F_CIB_UPDATE_RESULT); print_dot(output_format); if (current_cib != NULL) { rc = xml_apply_patchset(current_cib, diff, TRUE); switch (rc) { case -pcmk_err_diff_resync: case -pcmk_err_diff_failed: crm_notice("[%s] Patch aborted: %s (%d)", event, pcmk_strerror(rc), rc); free_xml(current_cib); current_cib = NULL; break; case pcmk_ok: cib_updated = TRUE; break; default: crm_notice("[%s] ABORTED: %s (%d)", event, pcmk_strerror(rc), rc); free_xml(current_cib); current_cib = NULL; } } if (current_cib == NULL) { crm_trace("Re-requesting the full cib"); cib->cmds->query(cib, NULL, ¤t_cib, cib_scope_local | cib_sync_call); } if (options.external_agent) { int format = 0; crm_element_value_int(diff, "format", &format); switch(format) { case 1: crm_diff_update_v1(event, msg); break; case 2: crm_diff_update_v2(event, msg); break; default: crm_err("Unknown patch format: %d", format); } } if (current_cib == NULL) { if(!stale) { print_as(output_format, "--- Stale data ---"); } stale = TRUE; return; } stale = FALSE; kick_refresh(cib_updated); } static gboolean mon_refresh_display(gpointer user_data) { xmlNode *cib_copy = copy_xml(current_cib); stonith_history_t *stonith_history = NULL; last_refresh = time(NULL); if (cli_config_update(&cib_copy, NULL, FALSE) == FALSE) { if (cib) { cib->cmds->signoff(cib); } print_as(output_format, "Upgrade failed: %s", pcmk_strerror(-pcmk_err_schema_validation)); if (output_format == mon_output_console) { sleep(2); } clean_up(CRM_EX_CONFIG); return FALSE; } /* get the stonith-history if there is evidence we need it */ while (is_set(options.mon_ops, mon_op_fence_history)) { if (st != NULL) { if (st->cmds->history(st, st_opt_sync_call, NULL, &stonith_history, 120)) { fprintf(stderr, "Critical: Unable to get stonith-history\n"); mon_cib_connection_destroy(NULL); } else { stonith_history = stonith__sort_history(stonith_history); if (is_not_set(options.mon_ops, mon_op_fence_full_history) && output_format != mon_output_xml) { stonith_history = reduce_stonith_history(stonith_history); } break; /* all other cases are errors */ } } else { fprintf(stderr, "Critical: No stonith-API\n"); } free_xml(cib_copy); print_as(output_format, "Reading stonith-history failed"); if (output_format == mon_output_console) { sleep(2); } return FALSE; } if (mon_data_set == NULL) { mon_data_set = pe_new_working_set(); CRM_ASSERT(mon_data_set != NULL); } mon_data_set->input = cib_copy; cluster_status(mon_data_set); /* Unpack constraints if any section will need them * (tickets may be referenced in constraints but not granted yet, * and bans need negative location constraints) */ if (show & (mon_show_bans | mon_show_tickets)) { xmlNode *cib_constraints = get_object_root(XML_CIB_TAG_CONSTRAINTS, mon_data_set->input); unpack_constraints(cib_constraints, mon_data_set); } switch (output_format) { case mon_output_html: case mon_output_cgi: if (print_html_status(out, output_format, mon_data_set, stonith_history, options.mon_ops, show, print_neg_location_prefix) != 0) { fprintf(stderr, "Critical: Unable to output html file\n"); clean_up(CRM_EX_CANTCREAT); return FALSE; } break; case mon_output_legacy_xml: case mon_output_xml: print_xml_status(out, mon_data_set, stonith_history, options.mon_ops, show, print_neg_location_prefix); break; case mon_output_monitor: - print_simple_status(mon_data_set, stonith_history, options.mon_ops, output_format); + print_simple_status(out, mon_data_set, stonith_history, options.mon_ops); if (is_set(options.mon_ops, mon_op_has_warnings)) { clean_up(MON_STATUS_WARN); return FALSE; } break; case mon_output_console: /* If curses is not enabled, this will just fall through to the plain * text case. */ #if CURSES_ENABLED blank_screen(); print_status(out, output_format, mon_data_set, stonith_history, options.mon_ops, show, print_neg_location_prefix); refresh(); break; #endif case mon_output_plain: print_status(out, output_format, mon_data_set, stonith_history, options.mon_ops, show, print_neg_location_prefix); break; case mon_output_unset: case mon_output_none: break; } stonith_history_free(stonith_history); stonith_history = NULL; pe_reset_working_set(mon_data_set); return TRUE; } static void mon_st_callback_event(stonith_t * st, stonith_event_t * e) { if (st->state == stonith_disconnected) { /* disconnect cib as well and have everything reconnect */ mon_cib_connection_destroy(NULL); } else if (options.external_agent) { char *desc = crm_strdup_printf("Operation %s requested by %s for peer %s: %s (ref=%s)", e->operation, e->origin, e->target, pcmk_strerror(e->result), e->id); send_custom_trap(e->target, NULL, e->operation, pcmk_ok, e->result, 0, desc); free(desc); } } static void kick_refresh(gboolean data_updated) { static int updates = 0; time_t now = time(NULL); if (data_updated) { updates++; } if(refresh_timer == NULL) { refresh_timer = mainloop_timer_add("refresh", 2000, FALSE, mon_trigger_refresh, NULL); } /* Refresh * - immediately if the last update was more than 5s ago * - every 10 cib-updates * - at most 2s after the last update */ if ((now - last_refresh) > (options.reconnect_msec / 1000)) { mainloop_set_trigger(refresh_trigger); mainloop_timer_stop(refresh_timer); updates = 0; } else if(updates >= 10) { mainloop_set_trigger(refresh_trigger); mainloop_timer_stop(refresh_timer); updates = 0; } else { mainloop_timer_start(refresh_timer); } } static void mon_st_callback_display(stonith_t * st, stonith_event_t * e) { if (st->state == stonith_disconnected) { /* disconnect cib as well and have everything reconnect */ mon_cib_connection_destroy(NULL); } else { print_dot(output_format); kick_refresh(TRUE); } } static void clean_up_connections(void) { if (cib != NULL) { cib->cmds->signoff(cib); cib_delete(cib); cib = NULL; } if (st != NULL) { if (st->state != stonith_disconnected) { st->cmds->remove_notification(st, T_STONITH_NOTIFY_DISCONNECT); st->cmds->remove_notification(st, T_STONITH_NOTIFY_FENCE); st->cmds->remove_notification(st, T_STONITH_NOTIFY_HISTORY); st->cmds->disconnect(st); } stonith_api_delete(st); st = NULL; } } static void handle_html_output(crm_exit_t exit_code) { xmlNodePtr html = NULL; out->finish(out, exit_code, false, (void **) &html); pcmk__html_add_header(html, "meta", "http-equiv", "refresh", "content", crm_itoa(options.reconnect_msec/1000), NULL); htmlDocDump(out->dest, html->doc); } /* * De-init ncurses, disconnect from the CIB manager, disconnect fencing, * deallocate memory and show usage-message if requested. * * We don't actually return, but nominally returning crm_exit_t allows a usage * like "return clean_up(exit_code);" which helps static analysis understand the * code flow. */ static crm_exit_t clean_up(crm_exit_t exit_code) { #if CURSES_ENABLED if (output_format == mon_output_console) { output_format = mon_output_plain; echo(); nocbreak(); endwin(); } #endif clean_up_connections(); free(options.pid_file); pe_free_working_set(mon_data_set); mon_data_set = NULL; if (exit_code == CRM_EX_USAGE) { if (output_format == mon_output_cgi) { fprintf(stdout, "Content-Type: text/plain\n" "Status: 500\n\n"); } else { fprintf(stderr, "%s", g_option_context_get_help(context, TRUE, NULL)); } } g_option_context_free(context); if (out != NULL) { if (output_format == mon_output_cgi || output_format == mon_output_html) { handle_html_output(exit_code); } else { out->finish(out, exit_code, true, NULL); } pcmk__output_free(out); } crm_exit(exit_code); }