Postfix restart automatically

[Mudasar]

Hi,

My zimbra is working fine for last 5 months, today suddenly i am getting pop3 error, and then it is fine. it is happened multiple time and still doing this. I have checked on internet but not find answer on it

I have also checked the physical memory it is almost 95% at that time. but not 100%.

Following is the error in maillog

-------------------------------------------------------
Feb 10 08:35:17 zm1 postfix/postfix-script[26805]: stopping the Postfix mail system
Feb 10 08:35:17 zm1 postfix/master[20361]: terminating on signal 15
Feb 10 08:35:25 zm1 zmmailboxdmgr[26979]: threaddump requested
Feb 10 08:35:25 zm1 zmmailboxdmgr[18625]: sending SIQUIT to mailboxd/JVM process 18626
Feb 10 08:35:27 zm1 zmmailboxdmgr[26986]: stop requested
Feb 10 08:35:27 zm1 zmmailboxdmgr[26986]: waiting for manager process 18625 to die
Feb 10 08:35:27 zm1 zmmailboxdmgr[18625]: shutdown requested, sending TERM signal to 18626
Feb 10 08:36:01 zm1 zmmailboxdmgr[18625]: mailboxd/JVM process exited (waitpid expected 18626 got 18626)
Feb 10 08:36:01 zm1 zmmailboxdmgr[18625]: manager woke up from wait on mailboxd/JVM with pid 18626
Feb 10 08:36:02 zm1 zmmailboxdmgr[26986]: manager process 18625 died, shutdown completed
Feb 10 08:36:04 zm1 zmmailboxdmgr[27283]: status requested
Feb 10 08:36:04 zm1 zmmailboxdmgr[27283]: file /opt/zimbra/log/zmmailboxd_manager.pid does not exist
Feb 10 08:36:04 zm1 zmmailboxdmgr[27283]: assuming no other instance is running
Feb 10 08:36:04 zm1 zmmailboxdmgr[27283]: file /opt/zimbra/log/zmmailboxd.pid does not exist
Feb 10 08:36:04 zm1 zmmailboxdmgr[27283]: assuming no other instance is running
Feb 10 08:36:04 zm1 zmmailboxdmgr[27283]: no manager process is running
Feb 10 08:37:09 zm1 zmmailboxdmgr[29828]: status requested
Feb 10 08:37:09 zm1 zmmailboxdmgr[29828]: file /opt/zimbra/log/zmmailboxd_manager.pid does not exist
Feb 10 08:37:09 zm1 zmmailboxdmgr[29828]: assuming no other instance is running
Feb 10 08:37:09 zm1 zmmailboxdmgr[29828]: file /opt/zimbra/log/zmmailboxd.pid does not exist
Feb 10 08:37:09 zm1 zmmailboxdmgr[29828]: assuming no other instance is running
Feb 10 08:37:09 zm1 zmmailboxdmgr[29828]: no manager process is running
Feb 10 08:37:09 zm1 zmmailboxdmgr[29835]: start requested
Feb 10 08:37:09 zm1 zmmailboxdmgr[29835]: checking if another instance of manager is already running
Feb 10 08:37:09 zm1 zmmailboxdmgr[29835]: file /opt/zimbra/log/zmmailboxd_manager.pid does not exist
Feb 10 08:37:09 zm1 zmmailboxdmgr[29835]: assuming no other instance is running
Feb 10 08:37:09 zm1 zmmailboxdmgr[29835]: file /opt/zimbra/log/zmmailboxd.pid does not exist
Feb 10 08:37:09 zm1 zmmailboxdmgr[29835]: assuming no other instance is running
Feb 10 08:37:09 zm1 zmmailboxdmgr[29836]: wrote manager pid 29836 to /opt/zimbra/log/zmmailboxd_manager.pid
Feb 10 08:37:09 zm1 zmmailboxdmgr[29837]: wrote java pid 29837 to /opt/zimbra/log/zmmailboxd_java.pid
Feb 10 08:37:09 zm1 zmmailboxdmgr[29836]: manager started mailboxd/JVM with pid 29837
Feb 10 08:37:40 zm1 postfix/postalias[30856]: warning: /etc/aliases, line 98: need name:value pair
Feb 10 08:37:41 zm1 postfix/postfix-script[30888]: warning: not owned by root: /opt/zimbra/postfix-2.4.7.5z/conf/amavis_client_whitelist
Feb 10 08:37:41 zm1 postfix/postfix-script[30889]: warning: not owned by root: /opt/zimbra/postfix-2.4.7.5z/conf/amavis_client_whitelist.db
Feb 10 08:37:41 zm1 postfix/postfix-script[30890]: warning: not owned by root: /opt/zimbra/postfix-2.4.7.5z/conf/main.cf
Feb 10 08:37:41 zm1 postfix/postfix-script[30892]: warning: group or other writable: /opt/zimbra/postfix-2.4.7.5z/conf/./amavis_client_whitelist
Feb 10 08:37:41 zm1 postfix/postfix-script[30893]: warning: group or other writable: /opt/zimbra/postfix-2.4.7.5z/conf/./amavis_client_whitelist.db
Feb 10 08:37:41 zm1 postfix/postfix-script[30907]: starting the Postfix mail system
--------------------------------------------------

Please this is urgent

Thanks

[uxbod]

Check /var/log/messages for any error messages. Does dmesg show anything at all ?

[Mudasar]

Hi,

I have attached the output of dmesg.

Previous messages are extract from maillog, i am getting following result when i am getting this problem.

[zimbra@ ~]$ zmcontrol status
Host zm.abc.com
mta Stopped
postfix is not running
saslauthd is not running
zmsaslauthdctl is not running

stats Stopped

Getting following messages in messages

Feb 10 08:37:41 zm1 saslauthd: auth_zimbra_init: zimbra_cert_check is off!
Feb 10 08:37:41 zm1 saslauthd: auth_zimbra_init: 1 auth urls initialized for round-robin

[Mudasar]

One more thing to add. There is 4 GB RAM in server, do you think it would be the possibility to crash the Postfix

I have attached the graphs of java heap.

[uxbod]

su - zimbra
zmlocalconfig | grep mailboxd_java_options
zmlocalconfig | grep mailboxd_java_heap_memory_percent

[Mudasar]

[zimbra@ root]$ zmlocalconfig | grep mailboxd_java_options
mailboxd_java_options = -client -XX:NewRatio=2 -Djava.awt.headless=true -XX:MaxPermSize=128m -XX:SoftRefLRUPolicyMSPerMB=1
[zimbra@ root]$ zmlocalconfig | grep mailboxd_java_heap_memory_percent
mailboxd_java_heap_memory_percent = 40

[uxbod]

Okay, a couple of things I would do :-

Modify /etc/sysctl.conf and add

vm.swappiness=0

and also

su - zimbra
zmlocalconfig -e mailboxd_java_heap_memory_percent=35

to tune down the amount of memory allocated to the java process. To activate all of these I would recommend giving your server a quick reboot.

[Mudasar]

Thnaks,

I have checked that machine swap is not used for cacheing, do you think it is nessery to change vm.swappiness=0

[root@zm1 ~]# free -m
total used free shared buffers cached
Mem: 3954 3850 104 0 263 1082
-/+ buffers/cache: 2503 1451
Swap: 8189 28 8160

[Bill Brock]

Originally Posted by uxbod

Okay, a couple of things I would do :-

Modify /etc/sysctl.conf and add

vm.swappiness=0

and also

su - zimbra
zmlocalconfig -e mailboxd_java_heap_memory_percent=35

to tune down the amount of memory allocated to the java process. To activate all of these I would recommend giving your server a quick reboot.

Why and when would adjustments like this need to be made? ie. Mine runs great with the defaults. Why would another server not do the same? Is there something in particular that would cause a need to stray from the defaults?

[uxbod]

Bill, what memory are you running ? profiles on different servers attribute aswell.