You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fix the issue that upgrade process may fail if the PD node is not available for longer than normal after restart (#1359, @AstroProfundis)
Fix incorrect MALLOC_CONF value for TiKV node, set prof_active to false (#1361#1369, @YangKeao)
Risk of this issue: Generating prof data for TiKV node with prof_active=true may cause high CPU systime usage in some circumstances, users need to regenerate startup scripts for TiKV nodes with tiup cluster reload <cluster-name> -R tikv to make the update applied
Fix the issue that the global log_dir not generated correctly for absolute paths (#1376, @lucklove)
Fix the issue that display command may report label mismatch warning if placement-rule is enabled (#1378, @lucklove)
Fix the issue that SELinux setting is incorrect when tiup-cluster tries to disable it with check --apply (#1383, @AstroProfundis)
Fix the issue that when scaling out instance on a host imported from tidb-ansible, the process may report error about monitor directory conflict (#1386, @lucklove)
Improvements
Allow scale in cluster when there is no TiSpark master node but have worker node in the topology (#1363, @AstroProfundis)