Veeam 6.5 slow direct SAN mode

[maverick964_uk]

We have a new physical proxy but the backups are slow. Veeam 6.5 and a HP proliant G8 server.

We see in the logs this:

Code: Select all

AgentMngr Starting Agent with lower priority, host "hostname.domain.org".

Could this be an issue and can we tune the priority?

At present direct SAN mode is slow and no quicker than a virtual proxy!

thanks

[Vitaliy S.]

Hi Mike,

Can you please post bottleneck stats for this job? What is the current job performance?

Thanks!

[dellock6]

Also, any reason you have not upgraded to v7?

[Gostev]

Process priority does not impact performance as long as there are CPU cycles available. In other words, if your server has no other roles but Veeam Backup Proxy, changing priority to any other value will not change anything. The cause of slow performance must be different, and as mentioned earlier, bottleneck stats should help you identify it.

[maverick964_uk]

hi guys, bottleneck stats dont seem to help however has anyone seen issues with performance based around Symantec Endpoint Protection manager? Or anything else that may slow the processing down...

I will get some bottleneck stats and post shortly. thanks!

[maverick964_uk]

Hello, are there a set of tools or processes to follow to try and assist with performance issues.

I am running v7 and the full backup is quite slow. The proxy is a physical so its using direct SAN.
I've shaved 1.5hrs off the full backup time by deinstalling the Symantec Endpoint Protection agent. I believe this impacts the backup but still unable to find the cause. It may be the virtual NiC driver SEPM installs...

The backup stats mainly show bottleneck of "Network" however with this backup, the proxy writes to a local disk.
If bottleneck is network then is this the comms between proxy/veeam/ESXi hosts/vcenter???

Its difficult to troubleshoot so would be useful to get some guidance. Logging levels, data in logfiles to check etc...

thanks!

[Gostev]

Hi, Mike

Network bottleneck means whatever connectivity there is between source data mover (which pulls data from production storage) and target date mover (which saves data to backup repository).

You cannot really do deep performance troubleshooting by yourself, you really should open a support case. Support will require you to enable a special performance logging mode, and then run your output through some data mining tools to make sense of it.

Thanks!

[maverick964_uk]

many thanks.