Re: ORA-4031 error and can't login to the db

thanks paul.

the shared pool is 600mb and the java pool is 150mb. i did end up killing the pmon process and restarting the database and the patch. everything is running fine now.

i always thought that the reserved pool was for the priveleged users to be able to log in to the db in a situation like this.  i can understand the 4031 error and shared pool sizing but i don't see how/why i couldn't log into the db without having to kill the process and restarting it. just wanting to be proactive on this, in case there is something wrong that i'm missing.

haroon

• Original message ----
  >Date: Thu, 23 Jun 2005 12:32:45 -0400
  >From: Paul Drake <bdbafh_at_gmail.com>
  >Subject: Re: ORA-4031 error and can't login to the db
  >To: haroon_at_qureshi.name
  >Cc: oracle-l_at_freelists.org
  >
  >On 6/23/05, Haroon A. Qureshi <haroon_at_qureshi.name> wrote:
  >> hello,
  >>
  >> i was applying the 11.5.10 maintenance patch for oracle 11i.
  >> the last phase of the patch is to recompile invalid objects.
  >> i recieved an ora-4031 trace file dump and everything is just
  >> hanging right now. i'm also not able to login to the
  >> database, either as sys or any user. it just sits there.
  >>
  >> has anyone seen this happen before or knows what's going on?
  >> the 4031 trace seems to have been generated when trying to
  >> recomile a view. from my understanding, the 4031 tends to
  >> happens when the SGA is too small and it can't allocate any
  >> more memory to it. but how can i get into the db to shut it
  >> down and resize everything? i don't know why i'm unable to
  >> log in and don't know how to get around it.
  >>
  >> help! :)
  >>
  >> thanks in advance,
  >> haroon
  >
  >Haroon,
  >
  >What was the size of the shared_pool when you were applying
  the patch?
  >As described in the readme for database software patchsets, the
  >shared_pool should be sized at minimum 150 MB (I prefer 160
  MB as its
  >a multiple of 16MB) when running catpatch.sql.
  >
  >I'd recommend killing the instance, starting up with a larger
  >shared_pool (and java_pool) size and re-running the patch.
  >
  >good luck.
  >
  >Paul
  >--
  >http://www.freelists.org/webpage/oracle-l
  -- http://www.freelists.org/webpage/oracle-l