search cancel

"Connection to Brightmail server DOWN" errors logged after update to SMG v10.6

book

Article ID: 163344

calendar_today

Updated On:

Products

Messaging Gateway

Issue/Introduction

Following the update to Messaging Gateway v10.6, some systems which have been in uninterrupted operation since version 9.5.4 or earlier may begin generating alerts indicating that the connection to the Brightmail server is down

bmclient_log
2016-05-25T09:00:06-04:00 (ERROR:2625.802383616): [9012] Connection to Brightmail server [127.0.0.1:41000 ID:0] DOWN. Will attempt to reconnect.
2016-05-25T09:00:06-04:00 (ERROR:2625.731887360): [9012] Connection to Brightmail server [127.0.0.1:41000 ID:0] DOWN. Will attempt to reconnect.
2016-05-25T09:00:06-04:00 (ERROR:2625.802383616): [9012] Connection to Brightmail server [127.0.0.1:41000 ID:0] DOWN. Will attempt to reconnect.
2016-05-25T09:00:16-04:00 (ERROR:2625.471852800): [9010] Unable to open new connection to Brightmail server [127.0.0.1:41000 ID:0].
2016-05-25T09:00:16-04:00 (ERROR:2625.471852800): [9012] Connection to Brightmail server [127.0.0.1:41000 ID:0] DOWN. Will attempt to reconnect.
2016-05-25T09:00:26-04:00 (ERROR:2625.731887360): [9010] Unable to open new connection to Brightmail server [127.0.0.1:41000 ID:0].
2016-05-25T09:00:26-04:00 (ERROR:2625.731887360): [9012] Connection to Brightmail server [127.0.0.1:41000 ID:0] DOWN. Will attempt to reconnect.

messages
2016 May 25 07:00:01 (notice) bmserver_memory_monitor: [8691] Monitoring process
2016 May 25 07:00:08 (notice) bmserver_memory_monitor: [8857] Restarting bmserver. vsize of 4386228 exceeds max allowed of 2900000
2016 May 25 07:30:01 (notice) bmserver_memory_monitor: [12929] Monitoring process
2016 May 25 07:30:08 (notice) bmserver_memory_monitor: [13096] Restarting bmserver. vsize of 4452292 exceeds max allowed of 2900000

cron
2016 May 25 04:30:02 (info) CROND: [20114] (root) CMD (sh /root/bmserver_monitor.sh)
2016 May 25 05:00:01 (info) CROND: [24188] (root) CMD (sh /root/bmserver_monitor.sh)
2016 May 25 05:30:01 (info) CROND: [28384] (root) CMD (sh /root/bmserver_monitor.sh)

 

Cause

On some very old Messaging Gateway systems a memory monitor cron job was installed to restart the Brightmail server process if the memory utilizatization approached the 32 bit process memory limit as a means of controlling memory utilization. Memory utilization by the Brightmail server process has long since been addressed but with the v10.6 release, Messaging Gateway moved to a full 64 bit process stack resulting in an increased memory footprint for some processes.

If the old memory monitor cron job is still running, it will restart the bmserver process every 30 minutes as the default process memory usage in v10.6 is greater than the upper limit on 32 bit process memory.

Resolution

This issue may be addressed in one of two ways:

  1. Reinstall the operating system on the affected appliance or virtual machine with the current 10.6 release (please see the appropriate KB articles for the osrestore process)
  2. Contact Symantec customer support so that a support representative can remove the memory monitor script from the system.