One of the media servers writing slow

[nasahasha]

Netbackup 7.5.0.1,Linux 5.5(All servers are on linux)

one of the 4 media server is writing slow oracle DB backups to AdvancedDisk,All four have same configurations and settings.I have gone through some points in tunning guide and settings are same for all 4.

Please tell me what else needs to be checked for performance issues?

I am ok with checking thing I have already checked.

In my troublshooting I found that /usr/openv/netbackup/db/config/shm directory do not exist on this media servers,Does creating this shm directory will impact speed?

how do I create this shm directory,what needs to be put in it?

Thank you in advance,

Naseer

[nasahasha]

Please let me know if I need to paste any info

[Marianne]

Is this media server backing up itself or clients on the network?

Do you have bptm log for this media server? If so, please rename to bptm.txt and post as file attachment.

[nasahasha]

Its backing up clients on network,I havent enabled bptm logs,I can enable it for 30mins or 1 hour if required.

I think below is the procedure to enable logs,I have 4G available in /usr hope this is enough.

create bptm directory in /usr/openv/netbackup/logs and add BPTM_VERBOSE = 5 to bp.conf,do I have to run bprdreq --rereadconfig?

[Marianne]

No need for verbose logs. Simply create bptm log folder and leave it. No need to restart anything on media server. Next job that starts will start logging.

You need at least one job from beginning to end to have anything meaningful in bptm log.

Are you backing up same clients to specific media servers all the time? What happens if you direct these clients to another media server?

Have you tried to troubleshoot outside of NBU by ftp'ing a large file from a client to the media server?

[nasahasha]

I have created bptm,no verbose,will post after 30 mins

It not same client all times,clients change as we have media server load balancing for SUG<backups are writing good on other media servers.

never tried troubleshoot outside of NBU,I can try to ftp a large file from any client,Please let me know how can I do it and what needs to be monitor.

[nasahasha]

Please find bptm log as attachment.There were no new filesytem backups triggred but few DB archive continous and neew database incr backup jobs were ran.

[Yasuhisa_Ishika · Accepted solution]

'waited and delay' counter of child bptm shows no delay - all zero, so the bottleneck point is somewhere between child bptm and client operations.
If same policy works file with other media servers, I doubt network confguration and performance. Try to ftp a large file between client and media servers, and compare the performance.
It is better to ask OS and network team to help you.

[nasahasha]

Thank you