Re: Thread 1 cannot allocate new log, sequence

On Oct 2, 11:29 am, ddf <orat..._at_msn.com> wrote:
> On Oct 2, 8:52 am, "odbio..._at_spamu.pl" <odbio..._at_spamu.pl> wrote:
>
> > Hi,
>
> > I have 2 "Thread 1 cannot allocate new log, sequence" messages in my oracle's alert log. This messeges were few days ago and didn't come back. Is it
> > problem. Should I do something with it?
>
> > Thanks in advance
>
> It might be, it might NOT be.
>
> Check your redo log size
>
> SELECT     group# "GROUP",
>         bytes "SIZE",
>         status "STATUS"
> FROM       v$log;
>
> and your redo log switch time
>
> SELECT ROUND(AVG(1440 * (b.first_time - a.first_time)), 0) "Log switch
> time - minutes"
> FROM v$loghist a, v$log b
> WHERE b.sequence# = a.sequence# + 1
> AND a.sequence# = (SELECT MAX(sequence#) FROM v$loghist)
> ORDER BY a.sequence#;
>
> You're probably running  that database in archivelog mode and the
> error was generated because of an archiving issue.  It could also be
> generated by a 'checkpoint not complete'  problem.  Transactional
> activity in your database may have hit a new 'high' at the point those
> error messages appeared, causing Oracle to fall behind on it's
> archiving of the older redo logs prior to re-use because the log
> switch time decreased dramatically.
>
> Of course without actually seeing your system and its activity level
> all of this is simply guesswork.
>
> David Fitzjarrell

I agree with David and Hans. Look at how often you switch redo logs then perform some combination of adding more online redo logs or make them bigger if you regularly see places where the online redo logs are switching within only a couple of minutes. It is generally fine if you have one short switch time every day as long as the other times are more reasonable. But it you have an extended period where you are constantly switching you need bigger logs.

HTH -- Mark D Powell -- Received on Fri Oct 03 2008 - 07:38:33 CDT