not really, this server have two cpu and 2Gb memory, the cpu load average is arround 30% and the memory used is 900Mb

Hi. we use mercury32 for a long time , now the 4.62 version and w2k3 as server

some times (two, three times each month) we experience temporal problem with a lot of mercury restarts...  allways the situation is solved deleting all files within the mailqueue directory..

somebody have ideas about how fix it???

here is some lines from loader.log (for today's drops)

08-11-04.1304: Normal operation restored - resetting counters.
08-11-04.1522: Restarting Mercury after apparent abnormal termination
08-11-04.1524: Restarting Mercury after apparent abnormal termination
08-11-04.1524: Restarting Mercury after apparent abnormal termination
08-11-04.1524: Restarting Mercury after apparent abnormal termination
08-11-04.1524: Restarting Mercury after apparent abnormal termination
08-11-04.1524: Restarting Mercury after apparent abnormal termination
08-11-04.1524: Restarting Mercury after apparent abnormal termination
08-11-04.1524: Too many abnormal terminations in time period - attempting recovery.
08-11-04.1524: Quarantined 1149 queue files in 'BAD-08-11-04.1524'.
08-11-04.1524: Recovery complete - attempting to restart Mercury.
08-11-04.1530: Restarting Mercury after apparent abnormal termination
08-11-04.1531: Restarting Mercury after apparent abnormal termination
08-11-04.1531: Restarting Mercury after apparent abnormal termination
08-11-04.1532: Restarting Mercury after apparent abnormal termination
08-11-04.1532: Restarting Mercury after apparent abnormal termination
08-11-04.1533: Restarting Mercury after apparent abnormal termination
08-11-04.1533: Restarting Mercury after apparent abnormal termination
08-11-04.1533: Abnormal terminations continued after attempted recovery - exiting.
08-11-04.1533: Mercury/32 Loader shutting down.
08-11-04.1544: Mercury/32 Loader Started
08-11-04.1545: Restarting Mercury after apparent abnormal termination
08-11-04.1545: Restarting Mercury after apparent abnormal termination
08-11-04.1546: Restarting Mercury after apparent abnormal termination
08-11-04.1546: Restarting Mercury after apparent abnormal termination
08-11-04.1546: Mercury/32 Loader shutting down.
08-11-04.1547: Mercury/32 Loader Started

What do the main Mercury logs say prior to a shutdown?

Have you recently changed or updated a resident anti-virus program on the server? 

Is there anything suspicious about any of the queue files?

no abnormal entries on logs before drops.

this behavior occur since upgrade to last version 4.62

this system only have instaled calmav antivirus managed from clamwall daemon

Is this server heavily loaded? I would do a deep diagnostic on the hardware - particularly memory. Only time I've seen this is with some faulty memory - taking it out cured the problem.