Sudden slow network mode backup

[itlad]

Hi all

We have a 3 node vmware cluster backed by a Gigabit iSCSI SAN and 2 x NAS arrays
Veeam is installed on one of the guest VM's and has been happily backing up VM's on the SAN/NAS for months
Typically we'd get transfer speeds of 60-120mb/s using either direct san access mode or network mode and that was plenty to backup approx 2TB on a Sunday.

However now direct san access backups are performing fine but backups of VM's on the NAS are now suddenly crawling at under 500kbps.
The source (NAS) is showing up as the bottleneck. As a test we moved one of the affected guest VM's to a new NAS that we are bringing into production and ran a test backup of that, noting that it failed over to network mode backup as expected and we got nearly 50mbps throughput.

So we concluded something is wrong with the original NAS and moved about 20 VM's to the new NAS to allow the other NAS to be fixed.
The VM's were migrated in a reasonable amount of time, no issues there.
BUT when we tried to back up the 20 VM's on the new NAS (That tested fine earlier) we've got the same problem again!!

Any ideas / suggestions would be appreciated....

[itlad]

********* RESOLVED *******

I changed the backup job proxy mode from automatic to the local LAN proxy
and then restarted the job and achieved around 50mbps
Veeam must have been using a remote Veeam proxy in one of our WAN sites
Not very clever of you Veeam!!

[Vitaliy S.]

Hi Colm, glad that you've nailed it! Can you please tell me what version of Veeam B&R are you using? Thanks.

[itlad]

Hi Colm, glad that you've nailed it! Can you please tell me what version of Veeam B&R are you using? Thanks.

Hi Vitaliy. That would be Version 6.0.0.153

Thanks...

[Vitaliy S.]

Oh, that explains it! I strongly recommend to upgrade to the most recent version (or at least to v6.1) as we have enhanced our proxy detection mechanism for exactly this kind of scenarios.

Here is a quote from What's New doc (v6.1):

Backup proxy server selection. Intelligent load balancing now uses a Network access mode backup proxy server only when there is no backup proxy server available that can process the VM in Direct SAN or Hot Add access mode. This prevents autodetect from picking a Network access mode backup proxy server from another site.

[Gostev]

Not very clever of you Veeam!!

It is not very clever to be 2 releases behind the current version either

By the way, in your case the correct quote from 6.1 What's New is this (and not what Vitaly has posted):

Subnet detection. Network access mode backup proxy servers are now chosen based on their subnet. The selection logic always tries to use a backup proxy server located on the same subnet as the processed host. If no such backup proxy server is found, the first least-busy Network access mode backup proxy server is selected, and a corresponding event is logged in the job’s session log.

[Vitaliy S.]

My bad, thanks Anton!

[itlad]

Have just found out this affects restores also.
Strange that it worked fine until now and suddenly this 'bug' has appeared....

Anyway - I accept that I need to upgrade to the latest version.
I will do that as soon as I've found a way to put more hours in the day

Thanks everyone....