Recovering data from different server without client

Create dummy client with same name as on server.

I assume this is NDMP data which means scanner won't populate index.  In that case you must use ssid restore.  And since this is filer data, it means you can restore it only back to same type of filer.

Thanks for the reply. I was using the SSID of the save set. By "same type of filer" does it have to be specifically the same model/hardware or any NetApp filer running a similar version of ONTAP?

Also, how would I get the clientid if the server that created the backup is no longer available?

EDIT: Found client id using mminfo -aVS on the media

I tried making a dummy client pointing to a different filer, then tried again and nsrndmp_recover said:

42795:nsrndmp_recover:ssid'3247694783': Performing recover from Non-NDMP type of device

Then did nothing; it didn't even load the tape (I'm using a Quantum SuperLoader 3, 16 slot library). Any ideas what's going on?

Have you added the dummy client with the same client ID?

I'm not sure it matters, but I'd do it that way.

/tony

Yes, I used the client ID of the original client from the save set I'm trying to recover. nsrndmp_recover is getting further than before but it seems to be hanging somewhere now - I'm not sure if the problem is trying to load the tape or what. NetWorker doesn't seem to recognise the volume label of the tape - the volume name doesn't appear in nsrjb or in the Management Console, though it is present in Media -> Volumes in the console, and it will list save sets there.

Inky,

I just located this article as well that may help:

How to recover a client without performing a recovery from the original Backup Server

EMC KB Article esg57949

Goal: How to recover a client without performing a recovery from the original Backup Server

Goal: How to rebuild a client without the use of the backup server

Fact: Windows 2000 sp 01

Fact: NetWorker Server for Windows

Fact: NetWorker Client for Unix

Fact: Solaris

Fix: - Install the same version of the same Operating System and same version of NetWorker (install the NetWorker backup server on the new machine which will become the old Client)

- If the new machine will have a different name, then you will need to edit the hosts file on the New Solaris machine and add an entry for the old machine with a fake ip address (10.10.10.1)

- Create the old machine as a client in NetWorker on the new Solaris machine. Make sure, you create it using the same client ID as it existed on the original Backup Server.

This can be found by running the following command from the Original NetWorker Server: mminfo -av -c client_name -r "client, clientid"

- On the New Solaris machine, run nwadmin, go to Client --> Client Setup and select old client (client that backed up the files, which you just created)

- Enter *@* in the Remote Access field and then Click Apply

- Scan the volume that contains the backup for the client.

- Now you will need to perform a Directed recovery by running nwrecover from the new Solaris machine

- Go to Change --> Client, in nwrecover, and select the source client (old machine) and click OK

- Mark the root directory to recover everything

- Then select Recover

This should do it. Goal: How to recover a client without performing a recovery from the original Backup Server

Goal: How to rebuild a client without the use of the backup server

Fact: Windows 2000 sp 01

Fact: NetWorker Server for Windows

Fact: NetWorker Client for Unix

Fact: Solaris

Fix: - Install the same version of the same Operating System and same version of NetWorker (install the NetWorker backup server on the new machine which will become the old Client)

- If the new machine will have a different name, then you will need to edit the hosts file on the New Solaris machine and add an entry for the old machine with a fake ip address (10.10.10.1)

- Create the old machine as a client in NetWorker on the new Solaris machine. Make sure, you create it using the same client ID as it existed on the original Backup Server.

This can be found by running the following command from the Original NetWorker Server: mminfo -av -c client_name -r "client, clientid"

- On the New Solaris machine, run nwadmin, go to Client --> Client Setup and select old client (client that backed up the files, which you just created)

- Enter *@* in the Remote Access field and then Click Apply

- Scan the volume that contains the backup for the client.

- Now you will need to perform a Directed recovery by running nwrecover from the new Solaris machine

- Go to Change --> Client, in nwrecover, and select the source client (old machine) and click OK

- Mark the root directory to recover everything

- Then select Recover

This should do it.

Hello Inky!

Please review the process for recovery of data outside the retention policy and see if this helps.  I beleive this will work for what you are attempting to do.

EMC KB article esg52699  located at  http://solutions.emc.com/EMCSolutionView.asp?id=esg52699&usertype=C   Please review the ENTIRE document BEFORE starting the process so you are fully aware of the steps being taken and the expected results.

Here is the information in the KB article mentioned:

Resolution
1. Recover the data as well as update the index and media database.

inky wrote: Thanks for the reply. I was using the SSID of the save set. By "same type of filer" does it have to be specifically the same model/hardware or any NetApp filer running a similar version of ONTAP? Also, how would I get the clientid if the server that created the backup is no longer available? EDIT: Found client id using mminfo -aVS on the media

I believe if it was NetApp it should be NetApp.  Similar ONTAP versions should be ok.

Looking down further, I see that message you get is that you are restoring data from nonNDMP drive.  Sounds as drive is not assigned to dummy filer (there is nothing dummy there actually, this is actual filer that will get data).  My understanding is that you saved data originally on filer with attached drives, right?  Now, to restore it, you need second filer with attached drives too.

Thanks Mark, I got it to recognise the media volume name after running that scanner command, however the uasm thing didn't work, I guess because it's an NDMP backup. The client is a NetApp Filer so I couldn't really use anything from your second reply.

@Hrvoje: I'm not sure exactly what you mean by "attached drives". This the filer I'm trying to restore to has a lot of disks and the same volume structure as the filer that the backup was made from.

My problem seems to still be that nsrndmp_recover does not understand the correct tape is mounted in the drive. In NetWorker Administration GUI if I go to Monitoring -> Log I get these two messages constantly when nsrndmp_recover is running and apparently doing nothing:

51302768466  event  media  Media notice: enable or check device(s); volume XFI918 needed

51302768514  event  media  Media notice: no matching devices on `server' for recover by client `dummy'

XFI918 is now being recognised by nsrjb, and NetWorker Administration (when I look at Devices -> Libraries) whereas before is was an "unknown" volume.

Inky,

OK - that is my fault - I should have caught that.

The usam command is for use of an auto recovery process and is mainly UNIX.

For Windows, use these commands - it will give you what you need

mminfo -av -c client_name -r "client, ssid, volume, sscreate"

scanner –i -S save_set_id device_name

You can use -iv  to get a versbose output during the scanner process - it dsepends on how much detail you like having during the process.

You can also take the mminfo command and output it to a .csv or .txt file to help ensure you have the information at hand if you need to run the scanner more than once.

NOTE: sscreate in the mminfo command will tell when the saveset was started.

Try this - let us know if this narrows it down for you.

Mark

I forgot to mention that the document I am getting this information from is the  NetWorker 7.4 Service Pack 1 Command Reference Guide, which is located on Powerlink at Support > Technical Documentation and Advisories > Software ~ J-O ~ Documentation > NetWorker Family > NetWorker > 7.4 & Service Packs.  It is a VERY useful guide for situations like this.

There is also the NetWorker 7.4 Multiplatform Disaster Recovery Guide located at the same location, on page 2.  I STRONGLY recommend you download and read through.

Mark

inky wrote: @Hrvoje: I'm not sure exactly what you mean by "attached drives". This the filer I'm trying to restore to has a lot of disks and the same volume structure as the filer that the backup was made from. My problem seems to still be that nsrndmp_recover does not understand the correct tape is mounted in the drive. In NetWorker Administration GUI if I go to Monitoring -> Log I get these two messages constantly when nsrndmp_recover is running and apparently doing nothing: 51302768466  event  media  Media notice: enable or check device(s); volume XFI918 needed 51302768514  event  media  Media notice: no matching devices on `server' for recover by client `dummy' XFI918 is now being recognised by nsrjb, and NetWorker Administration (when I look at Devices -> Libraries) whereas before is was an "unknown" volume.

Attached drives - drive were originally attached to filer.  This is different than when mixing ndmp stream with other data when drives are not attached directly to filer (dsa).  What you did originally will affect restore config you are doing now.

As for message, it indicates storage node field for client dummy to be in position to restore is not right one.  Obviously, if data was written directly by NetApp dump to tape, you need same config now and change storage node field of original client to the one you are restoring now (assuming this is direct attached drive config).

@Mark_Bellows: The NetApp Filer isn't Windows either, it uses a proprietary OS. I have looked at the Disaster Recovery Guide but it seemed to basically say "recreate everything you already had in place", and I don't have the resources to do this (this is just a test restore).

@Hrvoje Crvelin: I changed the storage node field to "nsrserverhost" (it was "dummy" for some reason), this is the same configuration as the filer that created the backup and the one I am performing the restore on, but it still seems to be having trouble:

51303141711 athena2 media Athena2: browsing

21303142441 event media Media critical event: waiting for LTO Ultrium-3 tape XFI918 on athena2

11303143041 event media Media alert event: waiting for LTO Ultrium-3 tape XFI918 on athena2

(Athena2 is the name of the backup server that I am running nsrndmp_recover on)

It looks to me that it still isn't recognising the tape in the drive, not sure though...

If original filer, from where data came, had storage node field nsrserverhost then I assume you used dsa to save the data (and not directly attached drives to Netapp filer).

Ok, so you initiate recover and then server says "ok, this data is on tape XFI918...I need it... give it to me".... and then this gets loaded or not?  Because you say that it does not recognize tape in drive which sounds as if tape is loaded in drive (on backup server).  When you do mminfo -q volume=XFI918 -r location, do you get answer to be library where tape is?  If not, do inventory.  If yes, and tape is not loaded, what happens if you load it manually?