I just noticed a detail that I don't like.
We had an error logged in the alert log. It informed me that details were in
db_ckpt_1234.log. Fine, but there was no such log in the bdump directory.
The dba who was here before me wrote a nice script to run every Sunday to clean up old trace files. What he didn't think of is that the trace files seem to be written per session, and that the checkpoint 'session' continues from instance start up.
His script came along and happily moved the trace file to a backup_bdump directory. Oracle couldn't find it and didn't write the details.
One more fix to make.
Subscribe to:
Post Comments (Atom)
4 comments:
Also depending on OS this may not actually free the space back to the filesystem. Tracking this down can be great fun, finding the process that is holding onto the file after you have moved/deleted the file, when the sysadmin won't let you install or run lsof and your left with the more arcane unix commands :)
I'm running on Solaris & AIX. I do hope that these 2 are releasing the space. I haven't looked into this problem. My list is long enough already! Thank you for the heads up though.
unrelated note:
Are you going to the training days with Jonathan Lewis in November?
let me share my experience with regard to the service of sql mdf restore, it automatically eliminates data corruption issues in selected databases
Post a Comment