Client Index Backup

[zadigadabop]

Hello,
I am wondering how to backup all the client indexes to just one tape so that in case of a total NetWorker server failure all I would need to do is restore the indexes from one source instead of doing a mmrecover on all tapes.

If I try to back up /networker/nsr/index/* it will get the client folders but no other sub folders or files.
I have tried all sorts of ASMs with no luck.

The reason is that we do some clients that are off the network for a while (laptops) and if I need to recover the indexes it might take tapes a couple of months back, which may be a lot of tapes and a real pain.
Thanks you in adavance,
Steve

 

[605]

The perfect solution is a separate pool which only filters by save sets. For this filed simply specify:

save sets: index
bootstrap


That's already it.

 

[zadigadabop]

Thanks you fot your reply.
Let me make sure I have this right, because what I set up did not work.

I added a new Media Pool. Under Selection Criteria > Save Sets I put in:
index
bootstrap

For the Group I added the NSR server as a client and told it to back up the location of the indexes. That did not work so I added a directive to specifically backup the location every time with this:
<</networker/nsr/index>>
+always: .
I tried other directives.
Any ideas?
Thanks again,
Steve

 

[605]

I guess you are somehow new to NetWorker. Don't make it too complicated - anything else is already there by default.

NetWorker by default will automatically backup the server's databases at the end of the group when the server's client data will be backed up. By default this is the group Default with the NW's client as the only one.

However, as this group is not enabled by default, and if this is the only group, NetWorker will run these backups whenever this group will be started. You can easily recognize these save sets in the report by running "mminfo -av".

You do not need to change anything for your client setup,
directives etc. in order to run those backups - you just use the new pool to send these backups to a specific media (pool).

The only thing you must ensure is that you start the group (with the savegrp command) and not run a manual backup (with the save command).

 

[ofladung]

We backup the CFI and the database to a separate tape. I created a pool with the two savesets and added my daily backup groups (incremental backup). The weekly backup (full) is not a part of this pool because we store the tapes offsite. It makes no sense to replace the database tape every day. Keep in mind that you backup text files with a very good compression rate. It takes a long time until such a tape is full.

 

[zadigadabop]

Thanks,
I am now backing up indexes to a seperate tape. I hope this will help if our NW servers hard drive ever goes out again.
I think one of my initial concerns is actually a error in our config, I can't back up the servers file system that contains the indexes. It backs up the folders but not any of the files.
After I made the pool for the indexes I can do an ad-hoc backup using nwbackup on the server but that still only gets some of the files, usually just older ones.
I am thinking this must a config problem. A call to EMC support did not help.
I have ALL selected for the Save Set of the Group so what else could it be?
Thanks
Steve

 

[Rif123]

There are local directives in the filesystem where the indexes are stored. These makes all backup skip the files and directories of the indexes. On Windows the files are called nsr.dir and on UNIX thay are called .nsr. You should not back up your indexes by running a backup of the NetWorkers file system. The index of a client is usually being backed up when you have completed a backup.

If you want to ONLY backup the indexes you should run savegrp -O from command line, have a look in the command reference for that. That will only backup the indexes for the clients in the particular group you also speicify.

 

[zadigadabop]

I learned something new today. Now why didn't EMC just point that out when I called, is pretty obvious now that I see that!
Thanks all for your help.
Steve