Sign Up
Log In
Log In
or
Sign Up
Places
All Projects
Status Monitor
Collapse sidebar
openSUSE:Step:FrontRunner
drbd-utils
0001-crm-fence-peer-fix-timeout-with-Pacemaker-...
Overview
Repositories
Revisions
Requests
Users
Attributes
Meta
File 0001-crm-fence-peer-fix-timeout-with-Pacemaker-2.1-milli-.patch of Package drbd-utils
From 8a28be74bc6efa93931c957e54c01abb18b984fe Mon Sep 17 00:00:00 2001 From: Lars Ellenberg <lars.ellenberg@linbit.com> Date: Wed, 12 Jan 2022 13:50:35 +0100 Subject: [PATCH] crm-fence-peer: fix timeout with Pacemaker 2.1: milli seconds vs seconds MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit crmadmin timeout was in milli seconds for <= 2.0.x, but became a TIMESPEC with default seconds in >= 2.1. Up to 2.0.4, atoi() was used, which effectively ignores "trailing garbage", so we could get away with always appending "ms". But with 2.0.5, it became g_option_context_parse G_OPTION_ARG_INT, which "Cannot parse integer value “200ms” for --timeout" :-| So grep the help message for "timeout.*milliseconds", and if not present, append an explicit "ms" unit. Also tolerate both ": ok" (2.1) and " (ok)" (older) when matching the output string of crmadmin -S. --- scripts/crm-fence-peer.9.sh | 24 +++++++++++++++++++++--- scripts/crm-fence-peer.sh | 24 +++++++++++++++++++++--- 2 files changed, 42 insertions(+), 6 deletions(-) diff --git a/scripts/crm-fence-peer.9.sh b/scripts/crm-fence-peer.9.sh index 36590bd8..c943bf9f 100755 --- a/scripts/crm-fence-peer.9.sh +++ b/scripts/crm-fence-peer.9.sh @@ -392,6 +392,20 @@ check_cluster_properties() crm_is_not_false ${stonith_enabled:-} && stonith_enabled=true || stonith_enabled=false } +setup_crm_timeout_unit_ms() +{ + # crmadmin timeout was in ms for <= 2.0.x, + # but became a TIMESPEC in >= 2.1. + # Up to 2.0.4, atoi() was used, which effectively ignores "trailing + # garbage", so we could get away with always appending "ms", but with + # 2.0.5, it became g_option_context_parse G_OPTION_ARG_INT, which + # "Cannot parse integer value “200ms” for --timeout" :-| + if crmadmin --help 2>&1 | grep -q -e "--timeout=.*in milliseconds"; then + crm_timeout_unit_ms="" + else + crm_timeout_unit_ms="ms" + fi +} # # In case this is a two-node cluster (still common with @@ -737,6 +751,7 @@ drbd_peer_fencing() local startup_fencing stonith_enabled check_cluster_properties + setup_crm_timeout_unit_ms if ! $had_constraint_on_entry ; then @@ -1075,14 +1090,17 @@ _check_peer_node_reachable() # it is obviously reachable. # # Do this only after we have been able to reach a DC above. - # Note: crmadmin timeout is in milli-seconds, and defaults to 30000 (30 seconds). + # Note: crmadmin timeout defaults to 30 seconds. + # # Our variable $cibtimeout should be in deci-seconds (see above) # (unless you use a very old version of pacemaker, so don't do that). # Convert deci-seconds to milli-seconds, and double it. + # See also setup_crm_timeout_unit_ms() above. + # if [[ $crmd = "online" ]] ; then local out - if out=$( crmadmin -t $(( cibtimeout * 200 )) -S $DRBD_PEER ) \ - && [[ $out = *"(ok)" ]]; then + if out=$( crmadmin -t $(( cibtimeout * 200 ))$crm_timeout_unit_ms -S $DRBD_PEER ) \ + && [[ $out = *@(": ok"|" (ok)") ]]; then peer_state="reachable" return fi diff --git a/scripts/crm-fence-peer.sh b/scripts/crm-fence-peer.sh index cb5deded..96786734 100755 --- a/scripts/crm-fence-peer.sh +++ b/scripts/crm-fence-peer.sh @@ -244,6 +244,20 @@ check_cluster_properties() crm_is_not_false $stonith_enabled && stonith_enabled=true || stonith_enabled=false } +setup_crm_timeout_unit_ms() +{ + # crmadmin timeout was in ms for <= 2.0.x, + # but became a TIMESPEC in >= 2.1. + # Up to 2.0.4, atoi() was used, which effectively ignores "trailing + # garbage", so we could get away with always appending "ms", but with + # 2.0.5, it became g_option_context_parse G_OPTION_ARG_INT, which + # "Cannot parse integer value “200ms” for --timeout" :-| + if crmadmin --help 2>&1 | grep -q -e "--timeout=.*in milliseconds"; then + crm_timeout_unit_ms="" + else + crm_timeout_unit_ms="ms" + fi +} # # In case this is a two-node cluster (still common with @@ -426,6 +440,7 @@ drbd_peer_fencing() local startup_fencing stonith_enabled check_cluster_properties + setup_crm_timeout_unit_ms if [[ -z $have_constraint ]] ; then # try to place it. @@ -718,14 +733,17 @@ check_peer_node_reachable() # it is obviously reachable. # # Do this only after we have been able to reach a DC above. - # Note: crmadmin timeout is in milli-seconds, and defaults to 30000 (30 seconds). + # Note: crmadmin timeout defaults to 30 seconds. + # # Our variable $cibtimeout should be in deci-seconds (see above) # (unless you use a very old version of pacemaker, so don't do that). # Convert deci-seconds to milli-seconds, and double it. + # See also setup_crm_timeout_unit_ms() above. + # if [[ $crmd = "online" ]] ; then local out - if out=$( crmadmin -t $(( cibtimeout * 200 )) -S $DRBD_PEER ) \ - && [[ $out = *"(ok)" ]]; then + if out=$( crmadmin -t $(( cibtimeout * 200 ))$crm_timeout_unit_ms -S $DRBD_PEER ) \ + && [[ $out = *@(": ok"|" (ok)") ]]; then peer_state="reachable" return fi -- 2.40.0
Locations
Projects
Search
Status Monitor
Help
OpenBuildService.org
Documentation
API Documentation
Code of Conduct
Contact
Support
@OBShq
Terms
openSUSE Build Service is sponsored by
The Open Build Service is an
openSUSE project
.
Sign Up
Log In
Places
Places
All Projects
Status Monitor
