Just to follow-up, I may have found a workaround to the lack of a db keepalive...
If I kill -HUP the authdaemon parent pid, it closes the db connections, albeit, not gracefully. Not sure what happens to any current queries...I suspect they are killed too, but I should get less failed authentications due to the db connections timing out. kill -HUP `cat /usr/local/share/sqwebmail/var/authdaemon/pid` I put the kill -HUP in the /etc/cron.hourly/sqwebmail-cron-cleancache before the call to cleancache.pl. The timeout is 1 hour in the firewall, so it is cutting it close to run once an hour, but hopefully it will due. And once it is launched for all users, maybe timeouts won't even be an issue. Please let me know if this is a bad idea. Thanks, Trevor On Wed, 20 Aug 2003, Trevor Astrope wrote: > Hello, > > Our database machine is behind a stateful firewall and I've been noticing > that the sqwebmail db connections are being timed out by the firewall. Is > it possible for sqwebmail to use a keepalive or ping mechanism to keep the > connection alive? Or could sqwebmail be configured to establish a new db > connection after a defined timeout (hardtimeout)? > > Thanks, > > Trevor > > Regards, Trevor Astrope [EMAIL PROTECTED]
