There were multiple cases when low default startup/shutdown systemd timeouts, namely mcs-dmlproc, mcs-workernode, result in a lost or corrupted extent map.
DMLproc is unable to complete a rollback and goes into a crash loop
ROLLBACK of a long lasting DML left cluster in unusable state
DMLProc failed to rollback a txn
mariadb services on column cluster nodes occasionally enters a restart loop
loadbrm - increase timeout
Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.