On Exadata dbnodes, we noticed that Clusterware 12.1.0.2 ora.cvu resource generating XML reports every 6 hours and occupied large amounts of space on /u01 filesystem. Perhaps this is the same on non-Exadata systems too, if you are running 12.1.0.2 Clusterware.
Upon looking into MOS ID (reports from ora.cvu resource using excessive disk space (Doc ID 1964482.1) it was learned that this is an expected behavior with 12.1.0.2 where ora.cvu invokes Clusterware Health check (CHC) every 6 hrs and stores the XML reports under $GI_HOME/cvu/baseline or $ORACLE_BASE/crsdata/@global/cvu/baseline' location. However, the CVU didn't clean-up the reports periodically which ultimately causes disk full.
As per the note, the bug is fixed as follows:
Upon looking into MOS ID (reports from ora.cvu resource using excessive disk space (Doc ID 1964482.1) it was learned that this is an expected behavior with 12.1.0.2 where ora.cvu invokes Clusterware Health check (CHC) every 6 hrs and stores the XML reports under $GI_HOME/cvu/baseline or $ORACLE_BASE/crsdata/@global/cvu/baseline' location. However, the CVU didn't clean-up the reports periodically which ultimately causes disk full.
As per the note, the bug is fixed as follows:
Bug 18143707 is fixed in 12.1.0.2
Bug 19703199 is fixed in 12.1.0.2.4 PSU
Bug 20177779 is fixed in 12.1.0.2.5 PSU
Quick workaround to this problem is to delete the files (xml) manually in order to release the space.
References:
reports from ora.cvu resource using excessive disk space (Doc ID 1964482.1)
ora.cvu Resource Reports not Purged (Doc ID 2054765.1)
12.1.0.2 Grid Infrastructure Patch Set Updates - List of Fixes in each GI PSU (Doc ID 1928853.1)
No comments:
Post a Comment