crm_report (8) Linux Manual Page

crm_report – Part of the Pacemaker cluster resource manager

Synopsis

crm_report –from date-time [–to date-time] [additional options] [dest]

Description

crm_report – Create a tarball containing everything needed when reporting cluster problems

Options

-V

increase verbosity

-v,–version

software version

–features

software features

-f, –from time

time prior to problems beginning: "YYYY-M-D H:M:S" (do not forget the quotes)

-t, –to time

time at which all problems were resolved: "YYYY-M-D H:M:S" (default: ‘now’)

-T, –cts test

CTS test or set of tests to extract

–cts-log

CTS master logfile

-n, –nodes nodes

node names for this cluster only needed if the cluster is not active on the current machine accepts both -n "a b" and -n a -n b

-l, –logfile file

log file to collect, normally this will be determined automatically

-p patt

additional regular expression to match variables to be removed (default: "passw.*")

-L patt

additional regular expression to match in log files for analysis (default: CRIT: ERROR:)

-M

collect only the logs specified by -l

-S, –single-node

single node operation; don’t try to start report collectors on other nodes

-c, –cluster type

force the cluster type (corosync,openais,heartbeat,logmaster)

-A, –openais

force the cluster type to be OpenAIS

-C, –corosync

force the cluster type to be CoroSync

-H, –heartbeat

force the cluster type to be Heartbeat

-u, –user user

ssh username for cluster nodes (default: root)

-D, –depth

search depth to use when attempting to locate files

-e, –rsh

specify the remote shell to use (default: ssh -T)

–sos-mode

use defaults suitable for being called by sosreport tool (behavior subject to change and not useful to end users)

–dest

a custom destination directory/file

dest

a custom destination directory/file

crm_report works best when run from a cluster node on a running cluster, but can be run from a stopped cluster node or a Pacemaker Remote node.

If neither –nodes nor –single-node is given, crm_report will guess the node list, but may have trouble detecting Pacemaker Remote nodes. Unless –single-node is given, the node names (whether specified by –nodes or detected automatically) must be resolvable and reachable via ssh.

Examples

crm_report -f "2011-12-14 13:05:00" unexplained-apache-failure crm_report -f 2011-12-14 -t 2011-12-15 something-that-took-multiple-days crm_report -f 13:05:00 -t 13:12:00 brief-outage