Thread: XLogWrite: had to create new log file

XLogWrite: had to create new log file

From
"Thomas T. Thai"
Date:
should i worry about these DEBUGs in my log file?

DEBUG:  MoveOfflineLogs: remove 00000004000000B6
DEBUG:  MoveOfflineLogs: remove 00000004000000B5
...
DEBUG:  XLogWrite: had to create new log file - you probably should do
checkpoints more often



Re: XLogWrite: had to create new log file

From
Peter Eisentraut
Date:
Thomas T. Thai writes:

> should i worry about these DEBUGs in my log file?
>
> DEBUG:  MoveOfflineLogs: remove 00000004000000B6
> DEBUG:  MoveOfflineLogs: remove 00000004000000B5
> ...
> DEBUG:  XLogWrite: had to create new log file - you probably should do
> checkpoints more often

No, but you probably should do checkpoints more often. ;-)

The way I understand it, a "checkpoint" is when the server flushes the old
log files (WAL files, not the message log).  It does that automatically
every few minutes, or when the log file has reached a certain size.  If
you get these messages then the latter case has happened, which is
presumably a bit of a performance hit.  Then if you do checkpoints more
frequently these won't happen because the log files are flushed when idle.

The documentation is admittedly lacking in that area.

--
Peter Eisentraut      peter_e@gmx.net       http://yi.org/peter-e/


Re: XLogWrite: had to create new log file

From
"Thomas T. Thai"
Date:
On Sat, 6 Jan 2001, Peter Eisentraut wrote:

> Thomas T. Thai writes:
>
> > should i worry about these DEBUGs in my log file?
> >
> > DEBUG:  MoveOfflineLogs: remove 00000004000000B6
> > DEBUG:  MoveOfflineLogs: remove 00000004000000B5
> > ...
> > DEBUG:  XLogWrite: had to create new log file - you probably should do
> > checkpoints more often
>
> No, but you probably should do checkpoints more often. ;-)
>
> The way I understand it, a "checkpoint" is when the server flushes the old
> log files (WAL files, not the message log).  It does that automatically
> every few minutes, or when the log file has reached a certain size.  If
> you get these messages then the latter case has happened, which is
> presumably a bit of a performance hit.  Then if you do checkpoints more
> frequently these won't happen because the log files are flushed when idle.
>
> The documentation is admittedly lacking in that area.

ok that made sense as i'm running mnogosearch with 10 indexers going. now,
here is a newbie question: how do i 'do chechpoints more often'? is there
something i can run from cron or something?