cancel
Showing results for 
Search instead for 
Did you mean: 

Logs in a clustered master

Juannillus
Level 4
Partner Accredited

Hello,

I am looking to get some information about the logging in NBU when configured in a clustered environment.

I have been looking in several docs, and it seems that I can't get the answer to my questions. The questions that I have are:

- Do I have to set up the log folders (Unified and Legacy) in a shared disk?

- If they are configured locally in each node, what happen to the logs when I have a failover? If the logs change to the new active node and I am performing some troubleshooting, then I need to check logs on both nodes?

I would to like to be pointed to the place where I can learn something more about logging in a clustered environment, how to set it up, which are the best practices... Does anyone know the docs to study this?

 

Kind regards,

 

Juan

1 ACCEPTED SOLUTION

Accepted Solutions

mph999
Level 6
Employee Accredited

 

Main NBU logs are located in ...

 

/usr/openv/neybackup/logs

/usr/openv/logs

These should not be on the shared disk - so you would check for the logs you need on the node that was active at the  time of any issue.

A very detailsed TN on logs in general.

http://www.symantec.com/docs/TECH75805

Martin

View solution in original post

9 REPLIES 9

mph999
Level 6
Employee Accredited

 

Main NBU logs are located in ...

 

/usr/openv/neybackup/logs

/usr/openv/logs

These should not be on the shared disk - so you would check for the logs you need on the node that was active at the  time of any issue.

A very detailsed TN on logs in general.

http://www.symantec.com/docs/TECH75805

Martin

Juannillus
Level 4
Partner Accredited

Hi Martin,

thanks for your prompt reply.

The logs not being on the shared disk, is there a reason for that?, what would happen if I set it up in that way?

I've been looking in the troubleshooting guide, in the NBU ClusteredMasterServer admin guide, in the TN that you point me to, and in some other websites, and I can't seem to find some guidelines on how to configure the logs in a cluster. have you heard of any such document?

 

Again thanks for your time.

Regards,

Juan

revarooo
Level 6
Employee

The reason for keeping the logs off the shared disk, if you have an issue that causes a problem, you will be able to tell which node was the problematic server.

 

mph999
Level 6
Employee Accredited

Persoanlly - I would leave them where they are ... ravaroo has just confirmed what I was about to say.

bpcd and vnetd are standalone processes - so trying to move these would be an issue.

You do not mention te OS - but it is possible to move the log location to the shared disk if you wish ...

Consider this thought ...

If an issue specific to the physical node causes a log entry, then the log fails across with the cluster you know have an entry that is not relevant to that node in the logs being written by that node.

I just asked one of the most experienced guys here what they would do - they would leave them where they are (but created in sepatae mount points).

Martin

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

I have seen a particular site where the logs were moved to the shared drive.
Due to some problem, Symantec Support asked to increase logging level to highest possible level for every conceivable type of log.
Within an hour the shared disk filled up, bringing down EMM database, causing service group to attempt failover. Understandably NBU was not able to start on 2nd node.
Symantec Engineer was blamed for bringing the system down...

So - DON'T move logs to the same share as the NBU databases.
If you REALLY want to, create a separate volume/mount point for logs and add to the cluster config.

 

mph999
Level 6
Employee Accredited

Morning Marianne <wave> ...

So, you have the advice there of Marianne (one of the most experienced NBU people probably anywhere ...), myself and a colleague of mine  ...

I would take this as a very very big hint, and leave things alone ...

Martin

Nicolai
Moderator
Moderator
Partner    VIP   

We got so tired of fetching log from different sides of the cluster  we decided to have one log destination. Furthermore extreme logging caused issue with local file space - wort case scenario is 1GB of log file per minute.

The log destination is a separate volume and disk group (we are on Linux and SFHA). The diskgroup and volume is part of the nbu_group.

/usr/openv/logs point to /logs/openv/

/usr/openv/netbackup/logs point to /logs/netbackup

Until this has been working perfectly - but with NBU 7.x where some of the daemons are running in standalone mode this may change.

mph999
Level 6
Employee Accredited

I think here there is no real answer, it is what you wish to do.  However, the bpcd and vnetd process being standalone will cause issues.

Martin

 

Juannillus
Level 4
Partner Accredited

Hello all,

sorry for the delay answering but I have been out of the office the last couple of days. It may be true that there isn't a specific answer to my question, but based on the answers we are gonna set the logs up, in two local, independent FS, instead of a unique shared FS.

Although the solution for me come from all the post together, I'll mark mph999 as the solution because he was the first one pointing me to the right path.

Thanks all for your time and your help, it's been usefull.

Regards,

Juan