Restoring index from 7.4.2 with 7.6.1

There have been several changes to the media database since 7.2, so I'm not too sure if that may present any issues.  Have you verified whether or not there are any index backups for that client from the specified time or soon after (or does this affect all indexes)?  If so one thing you can try is first running nsrck -L5 clientname and then running the nsrck -L7 again as there may be some corruption in the existing index.

It is true that there have been changes in the databases and indexes. However, this proces will still work as I have laid it out.

To be 100% sure, I contacted one of the Networker engineers and he replied that the commands and the sequence is the same regarless of the version of Networker. This is to allow for end-users who have older version of Networker to be able import data from the tapes.

These commands should provide you the information you need.

Mark

I know they should work but I have seen instances before where a big separation of versions has caused issues.  It would be useful if we could establish the existence of the indexes so that we can define the problem a little clearer.

Reto - at worst we can recover your data using a saveset recovery or scanner so we should not worry about not being able to recover the data; Mark has documented the scanner process above and saveset recovery is clearly explained in the admin guide so I won't repeat it here.  It may be an idea to start that process off while we investigate the problem with indexes in parallel as I am sure getting the actual backup data back is a higher priority.

Hi,

It's no recommended to upgrade NW from 7.2.x to 7.6.x (in fact, at NPG, this is not allowed) due to some DB structure changes. Maybe, these can be the cause of the issue.

Regards.

Claudio

First, I discussed this issue engineering. They assured me that the process that is being attempted (recovering data from a 7.2 tape into a 7.6 environment) IS possible and is done on a regular basis. The engineers keep in mind that there are customers who have data retained on take for an extended period (10 years or more), so they made the changes to ensure these processes would work regardless of the vesion used to actually back up the data.

Second, although it is not possible to upgrade directley from NW 7.2 to 7.6, it is possible to perform a stepped upgrade: 7.2 > 7.5; 7.5>7.6. If this done properly, it will work. One note: I would recommend that Engineering be involved in the inital process to ensure that the information moves over correctly to the new format of the database.

However, the customer is not concerned with an upgrade, but rather simply recovering data from an older version of Networker.

If there is concern about this process, as it seems to be a bit contriversial, I would siggest calling the support line and talking to a support specialist about the process to perform this recovery.

Mark

Thanks for all the answers.

First I made a mistake on the title. We updated or  server from NetWorker 7.4.2 to 7.6.1. Only the client is a UNIX-machine  with NetWorker 7.2 version.

We updated the server with the steps from 7.4.2 -> 7.5 -> 7.6.1.

Then I saw on the step from 7.4.1 to 7.5 that the indexes were updated. But only the indexes on the disk.

I have this Problem only when I load back indexes from the tape.

How can I conwert (or update) the indexes from the tape?

Now I will test the answer from Mark "How to recover data from outside of retention policy".

When you run the scanner command it is not recovering the index, it is reading all data on the data and rebuilding a new index from what it finds there.  Because you have to go through the tape and look at the entire saveset backup to rebuild the index it is a long procedure.

When you run nsrck you are recovering the index files; since these are normally relatively small it is a quick process.

You should be able to stop scanner from the command line using ctrl-c

So I made it. And it works!

I did it with Marks instructions.

First load the volume with the command:

nsrjb -lnv -S (slot#) -f (device_name)

(slot#) -> Number of the slot ex. 175, (device_name) ex. \\.\Tape0

Then I started the scanner command:

scanner -iv -c (client_name) (device_name)

Then it took 2 hours, then I had the index back.

When I change the browse time on the NetWorker User I can see the files. :-)

Is this the fastest command for this feature?

With nsrck -t "MM/DD/YYYY" -L7 clientname it has 5 minutes or less.

And an other question: how can I stop the scanner command on the command line? With [ctrl + c]?

Now I undestand. The scanner command creates a new indes from the data on the tape.

But I backuped the index from the client and I want to restore the existing index.

So I made a new Thread for this question.

https://community.emc.com/message/529014#529014