Comprehensive data protection for all workloads
Post Reply
toonv10
Novice
Posts: 3
Liked: never
Joined: Aug 30, 2018 6:00 pm
Full Name: Toon Vandendriessche
Contact:

StoreOnce Compacting

Post by toonv10 »

Hi,

One of our customers has the following configuration:
- Two StoreOnce appliances, one in Datacenter A one in Datacenter B
- One GateWay server in DTC-A.
- Backup Jobs to StoreOnce in DTC-A
- Backup Copy Jobs from StoreOnce DTC-A to StoreOnce DTC-B
- 100Mbps link between both DataCenters

Backup copy jobs were running fine until one job started a 'compacting' task. The Compacting task took more than 24h so all other jobs failed. During the compacting task our gateway server was constantly reading and writing data from/to the StoreOnce in DTC-B over the slow WAN link.
Compacting is disabled in the Backup Copy Job and can't be enabled (settings are greyed out).
We only saw this compacting task once on two different jobs.
Can somebody explain why this compacting happened?

I already created a support case: 03427487.
foggy
Veeam Software
Posts: 21069
Liked: 2115 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: StoreOnce Compacting

Post by foggy »

Hi Toon, this reminds me of this existing thread. Any chance you have some VMs that had been backed up for some time before but now are not backed up anymore though still present in the full backup file? In this case GFS retention executes a mechanism similar to compact operation, which is much slower due to not using auxiliary metadata. If you get rid of such VMs (enable deleted VMs retention, for example), you should stop seeing this behavior.

Also, do you have the 'Gateway server and StoreOnce are connected over WAN' check box enabled in the repository settings? This should decrease the amount of traffic over WAN.
toonv10
Novice
Posts: 3
Liked: never
Joined: Aug 30, 2018 6:00 pm
Full Name: Toon Vandendriessche
Contact:

Re: StoreOnce Compacting

Post by toonv10 »

Hi Foggy,

We didn't had deleted vm's and we also already enabled the "Gateway server and StoreOnce are connected over WAN" check box. Today I received an explanation from the support team:

We have been talking to QA team and here is some findings:
Indeed occasionally under some circumstances we do use "compact" instead of synthetic full backup and this is exactly the thing happened in your case.

1) Compact may happen in case of GFS + 1 (or more) failed VMs + HP StorceOnce
2) As we could see, there was an issue: donderdag 21 februari 2019 15:57:06

Resource not ready: Backup repository. And only 1 VM was processed fine. These circumstances triggered compact.
and the same (compact) for all the VMs that failed except the one that did not fail.

In order to avoid this issue you need to deploy another gateway closer to the target StoreOnce device.
[/I]


So it seems that when a job fails, the "compacting" task is triggerd for the failed vm's.
Thanks for your help! Great support from Veeam, as always :wink:
foggy
Veeam Software
Posts: 21069
Liked: 2115 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: StoreOnce Compacting

Post by foggy » 1 person likes this post

Yes, the reason for "compact" in this case is the same as in the mentioned case - the VM in question is not present in the incremental restore point following the full backup.
Post Reply

Who is online

Users browsing this forum: Google [Bot], Ivan239, Paul.Loewenkamp and 164 guests