EMM Service is restarting randomly multiple times a day!!! Please help.

[mohanty]

I have Netbackup EMM server installed on Windows 2003 standard server. The netbackup version is 7.1.0.2. Since couple of week the netbackup emm service starting randomly in a day. There is no message in event log that netbackup is stopped. I get only 3 message in event viewer: Netbackup EMM service terminated unexpectedly and then "Netbackup service is starting" and after few seconds "Netbackup service started". Please help me resolving this issue as it has been couple of weeks now and my backups/duplication jobs are failing. Attaching nbemm log. Thanks

[mohanty]

<<Attaching NBEMM log>>

[RamNagalla]

hi,

could you post the server.log file from the location \\installpath\Veritas\NetBackupDB\log\server.log

[mohanty]

Hi Nagalla,

Please find attached server.log file

[Dyneshia · Accepted solution]

1) What anti virus are you running ? If macafee be sure to make exceptions for NBU. http://www.symantec.com/docs/TECH56658

2) How much free space do you have where NBU is installed at ? 

  You should have at least 10% , otherwise NBU will stop NBU services to prevent data corruption.

3) Please update to 7.1.0.4 and test.

http://www.symantec.com/docs/TECH181132

[mohanty]

I am running Macafee and to test the issue we have uninstall it completely but after uninstallation the issue persists.

I have total 100gb drive out of which 30gb free where netbackup is installed.

Did anyone able to find any thing from NBEMM log which I have attached?

[mohanty]

The issue is still there. I have uninstalled Macfee completely and installed Microsoft Forefront on the server.

But no Luck.

[Mark_Solutions]

I see this in the log:

Transport_Cache_Manager::find_i, unable to locate a free connection

A possibility is this tech note (though an earlier version and not the exact issue) :

http://www.symantec.com/docs/TECH72244

Being a Corba type error take a look at this tech note which may help as you are getting these messages:

CORBA exception ObjectNotActive,29:EMMServer::DestroyByReference,

http://www.symantec.com/docs/TECH75227

Also please take a look at a thread I started as it could be a contributing factor:

https://www-secure.symantec.com/connect/forums/hung-parent-jobs

Currently I have one Master with the third setting set to 8192 - but do it a little at a time

but also make sure you have plenty of tcpip ports avialable on the Master Server:

Add these registry keys and reboot the Master:

HKLM\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters\
DWORD – TcpTimedWaitDelay  - Decimal Value of 30
DWORD – MaxUserPort – Decimal Value 65534

A lot here I know but it could be one or a combination of these factors

Start with the registry keys and Corba tech note - if the issue persists deal with the desktop heap - if still a problem look at the first link above

Keep us updated

[mohanty]

Hi Mark,

I checked in my Master server and found desktop Heap setting is 

Windows SharedSection=1024,20480,2048. So it is same as per your 3rd link.

As per 2nd link I didn't find registry setting DWORD – MaxUserPort – Decimal Value 65534. This need to be checked by me.

1st link I need to verify

Thanks Mark for your suggestion. I will look into the tech notes and let you know.

[Mark_Solutions]

OK - if you don't have the MaxUserPort setting then your Master only has about 1500 ports to play with - this can become easily exhausted with 7.5 at busy times - well worth doing.

[Dyneshia]

This is for 7.5 but still some good tuning info for Master server's in general.http://www.symantec.com/docs/TECH202840

[mohanty]

Issue Has been Resolved!!!

Thanks Mark for your suggestion .After applying 7.1.0.4 patch the issue got fixed.

Appriciate All for your valuable comments.

Regards--Sameer