Bug 1803729
| Summary: | mapr-zookeepr fails to start (probably related to systemd changes) on recently-updated CentOS/RHEL | ||
|---|---|---|---|
| Product: | Red Hat OpenStack | Reporter: | Luigi Toscano <ltoscano> |
| Component: | openstack-sahara | Assignee: | Nobody <nobody> |
| Status: | CLOSED EOL | QA Contact: | Luigi Toscano <ltoscano> |
| Severity: | high | Docs Contact: | Andy Stillman <astillma> |
| Priority: | high | ||
| Version: | 13.0 (Queens) | CC: | mimccune, ted.williams |
| Target Milestone: | --- | Keywords: | Triaged, ZStream |
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2023-07-10 17:28:00 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
EOL time for 15, but the issue should be visible on 13 too. OSP13 support officially ended on 27 June 2023 |
Description of problem: The mapr-zookeper service, which is initialized during the cluster startup, fails to start with no apparent explanation ("failed because a configured resource limit was exceeded.") This seems to be related a systemd change/fix which led to: https://github.com/systemd/systemd/issues/8085 In any case, the issue is described by this MapR knowledge base article: https://mapr.com/support/s/article/MapR-ver-6-x-systemd-incorrect-update While the title refers to 6.0, the workaround also works on 5.2 The following change should be applied to the zookeeper init script: (/opt/mapr/)initscripts/zookeeper -- 87 RUN_AS_CMD="su -s $MAPR_SHELL -p $MAPR_USER -c" ++ 87 RUN_AS_CMD="runuser -s $MAPR_SHELL -p $MAPR_USER -c" On a first look, fixing this can't be done by simply patching the executable on the image, because the affected script is extracted when the MapR RPMs are installed, which happens during the deployment. Version-Release number of selected component (if applicable): All mapr version when the image is RHEL- it depends on a operating system change.