ORA-00600: internal error code, arguments: [20533], [2230], [557557500]

I got this error at the end of an RMAN backup last night. Looking about on metalink, I see it as an unreproducible bug in older patch sets (bug 3473790). I see the workaround as recreating the control file.

One thing changed yesterday - I increased the RMAN retention policy from the default 7 to 30 days. Meant to do that when I created the db a couple months ago, but overlooked it. Only even noticed it because I was bringing up a standby from a backup more than a week old, which was not something I really ever expected to do, and my rcp-compressed-logs shell script works off of v$archived_log.

hp-ux 11i, 9.2.0.6, nocatalog.

Anyone else run into this? Simple commands that aren't recorded in the alert log and mess up the control file so RMAN blows up make me nervous. Any suggestions of what view into the controlfile might show something corrupt? SHOW ALL in RMAN looks fine.

I'm not sure if I'm even going to bother submitting this to metalink. I know they need to know, but am not convinced it will do anything other than make me annoyed at them, since I don't know that there is any communication from frontline to development about non-easily-reproducible issues. Besides "we're not going to work on something unless you reproduce it."

jg

--
@home.com is bogus.  To reply privately, change the home.com to cox.net.