|
From: Edward J S. <ed...@cr...> - 2005-06-29 12:38:48
|
Edward J Shornock wrote: > David Rees wrote: > >>On 6/28/05, Lionel Bouton <lio...@bo...> wrote: >> >> >>>One possible culprit could be the way SQLgrey handles the database >>>cleanups since 1.6.0 (the zombie process points to a forked child problem). >> >> >>FWIW, Since FC4 released perl-Net-Server 0.88 I haven't had any >>problems with sqlgrey locking up... >> > > > Hmm...I'm still on 0.87 yet. I've tried the "workaround" that Mr. Bouton > suggested, I'll post a follow-up if changing the "$self->fork_cleanup()" > calls to "$self->cleanup()" resolves it. If it does, I'll try upgrading > to Net::Server 0.88 as well. > Update: Reverting back to $self->cleanup() seemed to work, there weren't any lockups after 15 hours or so. I upgraded Net::Server to 0.88 and changed it back to $self->fork_cleanup(). I restarted sqlgrey in within a few hours the lockup happened. I've since changed back to $self->cleanup() in the code, maybe I got lucky before... I'll post another update if this *really* stops the lockups, after giving it a few days of operation. |