Re: Locking Question on OPS??

Although you are using two separate schema on the two nodes, you do not say that these are guaranteed to be stored in independent tablespaces.

In OPS, 'pinging' occurs between instances when both instances wish to update data in the same file - and if objects are in the same tablespace, then they may be in the same file.

Did you check the v$SESSION_wait view on the two separate instances, or v$sesstat (again on the two instances). v$file_ping may also help (if your version of Oracle supports it).

In version 8 there are also some v$DLM_xxx views which contain further information, and in fact version 8 supports GV$ views which cover all instances from one node if you have enabled parallel query slaves.

--

Jonathan Lewis
Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk

ccdicky wrote in message <80vfkd$76m1_at_imsp212.netvigator.com>...
>Hi there,
>
>Recently, I get a big question on an OPS system. My OPS database contains
>two schema. A process updating the tables on Schema #1 only is always
>running on Node A and triggered to update every minutes in everyday. An
>other daily process inserting data into the tables on Schema #2 is running
>on Node B. At first, they were running very well, but over time, the daily
>process on Node B is hung sometimes. And while the daily process is
>hanging, the update process on Node A also get hung until the daily process
>is killed.
>
>I queried several system views, including v$lock, v$system_wait .. on both
>nodes but cannot find any clue. Anyone can tell me what else system views
>can figure out what's wrong on the OPS system. In addition, is there any
to
>enable Oracle session trace while the process want to be traced is running?
>
>Any help would be appreicated. Thanks!
>
>Dicky
>
>
Received on Tue Nov 23 1999 - 11:28:06 CST