Path: dp-news.maxwell.syr.edu!spool.maxwell.syr.edu!drn.maxwell.syr.edu!news.maxwell.syr.edu!border1.nntp.dca.giganews.com!border2.nntp.dca.giganews.com!nntp.giganews.com!newshosting.com!nx02.iad01.newshosting.com!post02.iad01.newshosting.com!not-for-mail
From: netcomradeNSPAM@bookexchange.net (NetComrade)
Newsgroups: comp.databases.oracle.server
Subject: Re: dbwr and lgwr don't shutdown
Date: Mon, 04 Apr 2005 22:05:01 GMT
Message-ID: <4251b934.518572828@localhost>
References: <424c39d2.161882905@localhost>
X-Newsreader: Forte Free Agent 1.11/32.235
Organization: Acecape, Inc.
Cache-Post-Path: nntp.acecape.com!unknown@p78-65.acedsl.com
X-Cache: nntpcache 3.0.1 (see http://www.nntpcache.org/)
Lines: 32
X-Complaints-To: abuse@newshosting.com
Xref: dp-news.maxwell.syr.edu comp.databases.oracle.server:239613

On Thu, 31 Mar 2005 18:10:22 GMT, netcomradeNSPAM@bookexchange.net
(NetComrade) wrote:

>This is a weird issue.
>
>We have a machine (v880/solaris8/latest recommended patch cluster),
>where a particular Standard Edition 9.2.0.5 database shuts down
>'cleanly' yet leaves two processes behind, which we cannot kill unless
>reboot. The same database (storage attached to multiple machines) on a
>solaris7/e4500 machine comes up/shuts down fine. (i tiried to shutdown
>the rest with aborts, as you can see from the logs, didn't help)
>
>Any idea?
>here's a log from shutdown on v880

The issue has been narrowed down to DISK_ASYNCH_IO=true
When set to false, it works, however, disk performance degrades 4-5X
It has been further narrowed down to QIO feature of Veritas.
When QIO is turned off, shutdown also works, but performance still
degrades 4-5x.

Veritas analyzed savecores and suggested to upgrade to supported
version (3.5+) and see what happens..

We can live with performance degration for now, since this is an OLTP
db (99%+ memory reads)

We'll look for proper upgrade paths...

.......
We use Oracle 8.1.7.4 and 9.2.0.5 on Solaris 2.7 boxes
remove NSPAM to email