Backup Exec 2014 test Dedup

[levd]

Hi there,

On the moment im testing out Backup Exec 2014 with a deduplication job.
The job runs fine, starts and finishes with succes according to the Backup Exec console.

The strange thing is, when i look deeper on the network interfaces it looks like the backup is still running..? Can someone explain this behaviour?

Thanks,

Levd

[levd]

I found out the puredisk content router was running what created the data transfers.
Can someone explain why the job is completed succesfully but the puredisk content router is still sending or recieving data?

If i stop de puredisk content router the datatransfer is stopped, enabling this service again will intiate the data transfer again..
Would love to hear what this service is doing, the deduplication backup job is succesfully completed hours ago. I cant see or monitor this service in Backup Exec.

[levd]

anyone>?

[VJware]

Is a "verify" job running ?

Do you see a small amount of data transfer or a large one ?

[levd]

There not one job running, only the service puredisk content router, if this is stopped the data transfer stops.
Its a very large amount of data be uploaded.

Good to say, im backing up the server i installed Backup Exec on to a NAS with deduplication, all jobs finish fine, puredisk content router keeps running and uploads data from the NAS to the backup server lots of data.

[VJware]

Just to clarify. Is the DeDupe storage folder created on the BE server or the NAS ?

If not the local BE server, then the only supported connection methods for the DeDupe folders are FC and iSCSI. Is the NAS configured in such a manner ?

[levd]

The NAS is connected to the BE server with ISCSI, creating a local disk
In Backup Exec i made a deduplication storage with this local disk.

[VJware]

You may see lots of read/write activity on the DeDupe folder such a queue processing, garbage collection, compaction, space reclamation etc even when a backup isn't running and as such you 'll probably see such activity on the iscsi network.

[levd]

Ah ok, that will be it, thank you for explaining.
That must be it.

Levd

[levd]

Just talked with our network engineer, and he wonders if such a lot of network is really normal?
The below stats show the traffic, im running a dedup backup once a day that takes max 1,5 hours so the rest is generated by the puredisk service.

12/12  06.00

15/12  10.30

---------------

total 40.30 hours

approx. 60Mbps

450MB per minute, total about 1TB

Total backup data written on the deduplication device is just 375GB

Can you do anything with the storaged.log?

[Gurvinder]

can you go to the resource monitor. Post screenshot of Disk and Network Tab with all its subsections expanded and spoold.exe selected. That might give us some clue.

[levd]

Here is a screenshot from the resource monitor, the job just completed succesfully.

Better resolution: http://i61.tinypic.com/4py5cn.jpg

[Gurvinder]

BSISCHIPHOL, is that the media server or name if the NAS whose storage is mounted as ISCSI locally on Media Server. There are activities like CRQP (content router queue processing), Compaction, Rebasing, Garbage Collection which happen on the Dedupe folder. These activities can be checked in Storaged.log. It is possible that since the dedupe folder is created on a ISCSI LUN you would see network activity even after the backup.

In the network section of resource monitor if spoold does not have any connections with the remote server that means that the backup is complete. If the activity is just shown against the media server or NAS I would suggest its because of ISCSI connection. I would suggest to check the storaged.log once the backup is done as well.

[levd]

Bs1schiphol is the name of the backup server (also the server thats backuped)
But i wonder is this kind of data transfers normal for compaction, rebasing, garbage collection etc?

We choose too backup with deduplication because of lower data transfer rates, but if i see this it might be wise to skip deduplication and do normal full and incremental backups.

If i look in the resource monitor, i see this connections, im not using ipv6. Could this be an issue?

[Gurvinder]

I think it should be ok in that case. It is not using the network. I checked my test system and it is the same here. Loopback should not be a problem.

[levd]

Ok Gurvinder Rait, so i guess we need to accept this kind of network traffic after the job is finished if we want to use deduplication. Or just use normal backups if im right?

[Gurvinder · Accepted solution]

Yes, spoold would have worked on local disk if this was configured locally but since this is ISCSI you will see traffic in the background.

With normal backups you would loose out on deduplication but with a normal B2D there are no queue processing task etc. so network traffic should not be there.