Data ONTAP Discussions

Questions regarding Snapmirror & Disaster Recovery

We have recently finished installing a FAS2020 at our DR location and it is successfully snapmirrored with our home office. Now I'd like to see if anyone has had any experience with SAN failures. In our case we have a WP and a NY SAN. If for some reason our WP location goes up in smoke what are the steps to take to make sure the NY location comes online. I know that first the volumes would have to be unrestricted so that they are no longer read-only so queiscing the volumes would be first step. Are there any steps necessary after that? I would think that the steps after would depend on the scenario so I have provided the only two I could think of below:

Case 1: Connectivity would cease and therefore cut the NY SAN off from WP. This is best case I think.

Case 2: I think would be if there was bad data replicating from WP to NY. A little more difficult.

How do you handle something like bad data replication if it occurs? If there is a 9 minute lag between replication and bad data does get caught in the transfer would reverting to a previous snapshot be the only way? Any detailed comments are appreciated. Also feel free to point me to any online documents. I am not afraid of reading. Thanks.


Re: Questions regarding Snapmirror & Disaster Recovery

In the case of a disaster, you would break the Snapmirror relationship from the destination side and then it will be writable. Then it may be up to you to redirect your clients to the destination side. You may have to set up shares if you haven't done so already.

As far as "bad data", I'm not really sure what you mean by that. How you might handle this depends on what you mean by it.

Re: Questions regarding Snapmirror & Disaster Recovery

Thanks for the response.

That's pretty much what I thought as far as network connectivity loss to the other SAN. We run VMware in our environment and have our operating systems, data and exchange on different flexible volumes that are snapmirrored to NY SAN. So by bad data I mean anything such as an OS crash in our environment in WP (anything like a faulty driver, corrupt database, etc) replicating to the NYSAN. The only thing I can think of is pulling the last good snapshot before something like that happens. I was wondering if anyone had any experience with something like that happening and the steps they took to recover.

Re: Questions regarding Snapmirror & Disaster Recovery

Snapmirror should only replicate bad data if there was bad data on the source, it should not create bad data. If you know what the bad data specifically you can recover that specific data from a previous snapshot. If you don't know at all and can handle it, you can rever an flexvol back to a given snapshot.

Re: Questions regarding Snapmirror & Disaster Recovery

Not sure if you are using iscsi/san or cifs/nfs. If it's cifs/nfs you don't need to do so much. If using iscsi/fcp/san here is what i briefly recommend.

Since I gather you are using vmware , I will briefly describe what I have done to take care of your problems. Both vmware and netapp have perl api's.

You can quiesce a vm by snapshoting the vm, while that is done you create a snapshot on the netapp (i rotate 4 of them) - so now you know that you have 4 good snapshots which you can use. Then on the otherside you can clone the lun and present that lun. If you want further explanation let me know and I'll go into a little more detail