This content has been marked as final. Show 8 replies
Did you check the metalink (M.O.S...) doc: ID 1326983.1... (some bugs are corrected in 220.127.116.11.. and you are in 18.104.22.168)...1 person found this helpful
What are your values of
for your home?
As suggested, check out the retention policies. Here is a little more information on the [url http://www.ora00600.com/scripts/databaseconfig/ADRCI.html]ADRCI and how to set those policies.1 person found this helpful
Unfortunately, I've no access to metalink for the moment (my company just started using Oracle databases).
These are the default values that I haven't yet modified :
SHORTP_POLICY : 720
LONGP_POLICY : 8760
SHORTP_POLICY : 720 --> 30 days --> means will run once in a month
LONGP_POLICY : 8760 --> 365 days --> means will run once in a year
That's strange because if I set the two retention policies to one hour and wait, the mmon process doesnt clean the data.
But if I execute the purge command without parameter (using the default retention policies), the files are cleaned up.
Indeed, it looks like a bug. I'll check on metalink (when account available) for more information.
Thanks for the help.
I have found the same to be true. Even if you set your policies to something silly like 1 hour, it won't necessary perform an auto-purge.
Another problem is that while you can auto-purge your Health Monitor runs, this only purges records from the v$hm_runs view and leaves associated records in the v$hm_findings and v$hm_recommendations as orphans in those table.
Oracle's only fix, according to an SR I opened, was to 'DROP_HM_SCHEMA' - an undocumented function.
Also, it doesn't housekeep the text versions of the alert and listener logs - you have to manually housekeep those yourself, which is quite ridiculous.
I spent a lot of time over the last couple of weeks on this and I consider it something that's not really ready for prime-time (in the same manner ASMCMD was awful a couple of releases back). Has potential if it can become a 'one-stop shop' for housekeeping logs, though. I'd also like to see it incorporate auditing housekeeping.
Ok, I've found something on Metalink : DocID 9500575 (Bug 9500575 : 11G LISTENER LOGS LOG_XYZ.XML ARE NOT AUTOMATICALLY PURGED)
It seems that it's only working for rdbms directory and not for listener files, for the moment...
Wow. It's even less use than advertized :)