Hello Sam,
-----Mensaje original-----
De: Sam Varshavchik [mailto:[EMAIL PROTECTED]]
> In any case: is this a normal behaviour?:
>
> [EMAIL PROTECTED] log]$ grep "SHUTDOWN:" maillog
> May 26 10:35:57 m3lnxsva01 courierd: SHUTDOWN: Stopping...
> May 26 11:20:06 m3lnxsva01 courierd: SHUTDOWN: Restarting...
> May 26 11:45:39 m3lnxsva01 courierd: SHUTDOWN: Restarting...
> May 26 11:48:58 m3lnxsva01 courierd: SHUTDOWN: Restarting...
> ...
This is not normal.
> i.e., so many restarts so frequent? I see no "Shutdown: respawn limit
> reached" in the last 2 days. Why is it happening? This maillog started the
Perhaps because you never had a 20 minute interval where no mail was coming
through the system. Respawning occurs only when the system has been idle
for 20 minutes, and not doing anything, or at respawnhi setting, whichever
comes first.
In any case, you have some external script or process that runs the "courier
restart" command. And, it also probably runs "courierldapaliasd restart",
which exposes the leak bug.
One particular process that runs "courier restart" would be webadmin. If
you were messing around with various configuration settings in webadmin, at
that particular time, you'd see this in the logs, which would be normal.
However webadmin only invokes courierldapaliasd if it changes any LDAP
configuration data.
You should apply the patch in my previous message to fix the original bug,
then identify what exactly causes courier restart to be invoked so
frequently.
Hello,
I have applied the patch this morning, and increased the number of LDAPALIAS processes to 10. But the restarting keep on going.
We are not using webadmin. And we are not running any script, at least not consciously.
It seems an internal issue: here is an extract of
# grep -A 1 "SHUTDOWN" maillog
May 28 12:44:22 m3lnxsva01 courierd: SHUTDOWN: Restarting...
May 28 12:44:22 m3lnxsva01 courierd: Waiting. shutdown time=none, wakeup time=Wed May 28 12:44:42 2003, queuedelivering=130, inprogress=5
--
May 28 12:51:52 m3lnxsva01 courierd: SHUTDOWN: Restarting...
May 28 12:51:52 m3lnxsva01 courierd: Waiting. shutdown time=none, wakeup time=Wed May 28 12:52:12 2003, queuedelivering=129, inprogress=2
--
May 28 12:57:38 m3lnxsva01 courierd: SHUTDOWN: Restarting...
May 28 12:57:38 m3lnxsva01 courierd: Waiting. shutdown time=none, wakeup time=Wed May 28 12:59:23 2003, queuedelivering=130, inprogress=6
--
May 28 13:08:12 m3lnxsva01 courierd: SHUTDOWN: Restarting...
May 28 13:08:12 m3lnxsva01 courierd: Waiting. shutdown time=none, wakeup time=Wed May 28 13:08:31 2003, queuedelivering=137, inprogress=7
--
May 28 13:36:50 m3lnxsva01 courierd: SHUTDOWN: Restarting...
May 28 13:36:50 m3lnxsva01 courierd: Waiting. shutdown time=none, wakeup time=Wed May 28 13:36:52 2003, queuedelivering=128, inprogress=2
--
May 28 15:16:14 m3lnxsva01 courierd: SHUTDOWN: Restarting...
May 28 15:16:14 m3lnxsva01 courierd: Waiting. shutdown time=none, wakeup time=Wed May 28 15:16:30 2003, queuedelivering=136, inprogress=3
--
May 28 15:22:09 m3lnxsva01 courierd: SHUTDOWN: Restarting...
May 28 15:22:09 m3lnxsva01 courierd: Waiting. shutdown time=none, wakeup time=Wed May 28 15:24:38 2003, queuedelivering=138, inprogress=12
As you can see, courierd is planning "wake times" very often. Why is it?
