Hi,
In 12.1 on IBM/AIX over here we have a DB where we had set db_recovery_file_dest_size
to a small 10GB: last night this FRA diskgroup exploded when the DB tried to dump another
archived log file with:
Unable to create archive log file '+FRA'
Thu Dec 26 22:01:34 2019
Errors in file /u01/<diag directory>/db_arc0_39125466.trc:
ORA-19504: failed to create file "+FRA"
ORA-17502: ksfdcre:4 Failed to create file +FRA
ORA-15041: diskgroup "FRA" space exhausted
ARC0: Error 19504 Creating archive log file to '+FRA'
(the DB dumped 200GB of arc log files therein -sic !-), and it caused disastrous unavailability.
We didn't even get the usual ORA-19815: WARNING: db_recovery_file_dest_size of X bytes is <pct>% used
when space occupancy was getting worse and worse.
MOS or Google returned me nothing when searching for db_recovery_file_dest_size + "ignored"...
We've never seen that before (db recovery file dest size ignored, no warning, no error whatsoever,
and eventually the diskgroup exploding and everything freezing on this server (even ASM instance
was unavailable thru SQL*Plus...).
Any idea why the alert log didn't show any alert ? Someone's seen that before perhaps, are there
conditions (combination with another parameter value... ?) that could elicit this crash ?
PS: alert log also shows lines "minact-scn: useg scan erroring out with error e:12751" - no idea what
that means.
Thanks.
Regards,
Seb