2 Iron

Disaster Recovery OSSR issue

I have a Windows 2008 R2 client and I am using on the backup server and the client.  My F:\ drive on the client is approx 650 Gb.  When a backup run, it backs up all the drives and OSSr-C, OSSR_A and OSSR_F.  OSSR_F is also the same size as the original F:\ drive.  Questions:

Why does OSSR_F do a full backup everytime a backup is run, full or incremental.

What can I do to prevent OSSR_F from running every time as it is taking up valuable space on a tape and it runs for hours.

What am I missing?



0 Kudos
2 Replies
2 Iron

Re: Disaster Recovery OSSR issue

Hi Karim,

This is most likely due to "File Server Resource Manger" Role being installed and "Storage Monitoring" enabled for F: drive which as a result FSRM writer is running.

Writer name: 'FSRM Writer'

Wrter Id: {12ce4370-5bb7-4c58-a76a-e5d5097e3674}

You can get more details by running Diskshadow --> list writers detailed or by running snapviewer.

Please open a support ticket and ask for the patches from NW134616, your version of NetWorker matches the binaries in the escalation.

Regards, Parviz...

0 Kudos
2 Iron

Re: Disaster Recovery OSSR issue

Not quite sure about FSRM. Here's what I think might be the case.

OSSR_F is a "hidden" part of the DISASTER RECOVERY: saveset, that is being saved with Saveset ALL on all clients running 7.6.2.x. Same goes for OSSR_C. These are the savesets that are being used when the client machine needs to be disaster recovered using ISO image DR (you have seen changes in Disaster recovery procedure, right?).

There are two aspects of the problem here:

- why is OSSR_F being saved for Disaster recovery as a block backup, when a "normal" restore of F: is usually sufficient - question then actually being why do we need OSSR_F at all, when we have F: to recover from?

- why is NetWorker saving OSSR_F in full, when most of the files on F: are probably not changed on daily basis?

For the first problem answer is simple: because Microsoft said so. Starting with Vista Microsoft has started "identifying" critical system disks, that need to be saved with backup, in order to be sure that the Disaster Recovery is possible. Unfortunately, Microsoft identifies disk as a critical one, as soon as you have at least one registered service running on that drive. For instance, your F: might be 2TB data dump, that has a one tiny registered Apache service running from F: - and because of that entire F: disk is marked as critical, and needs to be saved in full.

For the second problem answer is simple: because EMC implemented it like that. I agree that it would be much smarter to avoid saving same things in F: and OSSR_F: twice, and use for instance something like internal pointers in Index area to distinguish things needed for DR from "plain" files, but so far it has not been implemented like that. I have heard plans for this to be improved in next big NetWorker release - however we will have to wait and see if this is the case.

To sum it up: why always full with OSSR_F - because it has been implemented like that, EMC suggests splitting saveset ALL to two client definitions where one will save disks, other will save DISASTER RECOVERY:, and What can you do to prevent this from happening - you can use VSS:DISASTER_RECOVERY=off in Save Operations to switch DR backups off - but beware, you might end up without functional disaster recovery if you decide to turn it off completely (no DISASTER RECOVERY: backups at all).

best regards,


0 Kudos