SAP Backup Failed - error 12 - Keep Alive & Lock File in LOG

[Georges_N]

Hello

I hope my title had enough keywords - here's the issue:

Windows Server 2003 x64

NBU 7.1

SAP Backups

.UTL file has the correct directories for SAP backups

Backup begins at 10pm and crashes after 4, 5 or 6 hours (usual full backup is 6-ish hours)

Same error everytime, only on this server (we have dozens of SAP servers)

6/26/2013 4:09:07 AM - Error bpbrm(pid=11876) could not write Keep Alive to database client     

(this repeats itself every 30-60 seconds for about 3 minutes)

6/26/2013 4:11:47 AM - Critical bpbrm(pid=11876) from client *********: FTL - Lock file ?????????????????????????????????????????????????k is unlocked

I starred (***) the client name for security purposes, but those question marks are actually in the NBU log window.

It's followed by:

6/26/2013 4:11:47 AM - Error bpbrm(pid=11876) could not write Keep Alive to database client     

6/26/2013 4:11:47 AM - Critical bpbrm(pid=11876) from client *******: FTL - SAP BEGIN alter tablespace failed   

6/26/2013 4:11:47 AM - Error bpbrm(pid=11876) could not write Keep Alive to database client     

6/26/2013 4:11:47 AM - Critical bpbrm(pid=11876) from client *******: FTL - Backup operation aborted!     

6/26/2013 4:11:47 AM - Info bpbkar32(pid=6504) done. status: 12

It started about 6 weeks ago. We switched backup nights (from one day to another) and still get an error - we checked the Windows tasks to see what could be interfering but nothing... also strangely nothing in the Event Log unless we're looking at the wrong place (Server 2003 is old school...)

I saw one case on here that was "resolved" because the person had checked his BACKINT and UTL file but no precision on what was checked exactly.

We uninstalled/reinstalled the NBU client as well, no effect.

Also checked http://www.symantec.com/business/support/index?page=content&id=TECH53405 - but this isn't our case.

Any ideas?

Thanks in advance.

[Nicolai]

Firewall between client and master/media ?

If yes - you need to configure TCP_KEEPALIVE_INTERVAL. 

A small howto is listed in the last section of http://www.symantec.com/docs/TECH125896

[Georges_N]

Internal network - firewall is not the case

I checked out your link but we actually don't have those in the registry, unless they must be created?

Thanks

[Nicolai]

Yes - They must be created. If do not have any firewalls I don't think "TCP keep alive" will help, but it does not hurt to set this parameter.

http://technet.microsoft.com/en-us/library/cc739819(v=ws.10).aspx

[Georges_N]

What ## would you recommend for a Windows 2003 x64 server? (in milliseconds)... Since our current registry doesn't have that key, we have to create it but also select a range...

Thank you

[Nicolai]

Sorry - can't help you. I am on Linux.

But take a look at:

https://www-secure.symantec.com/connect/articles/tuning-windows-2003-and-2008-symantec-netbackup

I don't think Andres has that setting covered.

[Georges_N]

Thanks guys, 

We have the same issue as http://www.symantec.com/business/support/index?page=content&id=TECH139572 except the solution did not help (increase logging? ...)

I'm curious if Symantec has more on that issue above.

[Georges_N · Accepted solution]

Hi everyone

little update - our BASIS team was using a powershell script to launch the batch jobs but the script itself seems flawed - if we launch the jobs using a modified vb script instead, it never crashes at the end, or at any time really.

We will be looking into it with Microsoft, thanks for your help!!

TL;DR --> Powershell is causing the SAP databases to unlock during backup, causing failure no matter what the timeout is set to.