Re: LGWR and DBWR
Date: Thu, 28 Feb 2008 00:52:28 GMT
"Jonathan Lewis" <jonathan_at_jlcomp.demon.co.uk> wrote in message
> "Dereck L. Dietz" <dietzdl_at_ameritech.net> wrote in message
>> Oracle 10.2.0.3.0
>> Windows 2003 Server
>> Is it possible for improperly sized redo log files to be overwhelmed to
>> the point where it will cause a process which is performing a lot of
>> writes to multiple tables to stop processing without any error or reason
>> in a log file?
> There is a flaw in that question.
> What do you mean by the process "stopped" ?
> Do you really mean that the process seems to have
> stopped because it takes much longer than you expect ?
> Have you checked v$session_wait and v$session_event
> for the process to see where it is accumulating time (if
> it is still alive).
> If the log files are very small and lgwr and dbwr are having
> trouble keeping up with the workload, the session might
> spend a lot of time waiting on
> log buffer space
> log file sync
> write complete wait
> and various waits relating to checkpoint incomplete (but,
> as another poster pointed out, these would be recorded
> in the alert log).
> Jonathan Lewis
> Author: Cost Based Oracle: Fundamentals
> The Co-operative Oracle Users' FAQ
This is part of my continuing Disappearing Session Saga. By "stops" I mean the session/process just disappears abruptly well before it should be finished. No error displayed through any program exception and as far as I can tell from what log files the off-site DBA sent and from what he has said nothing in the logs either. Received on Wed Feb 27 2008 - 18:52:28 CST