joka2
2 Iron

Special group for index backups

Jump to solution

Hi all,

We have a NetWorker 7.6.1 datazone. When a group has finished backing up a client, the group initates a backup of the index for every client as it finishes it's backup to a index pool. That way we keep the backup of the indexes in a different pool from the client backups. This requres that we have a tape device ready at all times to mount a tape from the index pool.

My question is:

would it be  any worse if the index backup was not backed up right after the client backup finises? My idea is to have a special group called index backup containing all the clients and have it run in the morning after all the client backups finish?

Best regards,

Johannes

0 Kudos
1 Solution

Accepted Solutions
masonb
2 Iron

Re: Special group for index backups

Jump to solution

Johannes,

Yes it is fine for Production environment also, I just mentioned DR as an aside to show people are going what you suggest. It will just limit the volumes you would have to select from for any days index recovery. Depending upon your total index size it could take some time to fill a volume to capacity so this sometimes introduces a single point of failure to the environment but cloning these backups can alleviate this.

Whether DR or production the important thing is they are secured and we can recover quickly - this approach should satisfy whether recovering all or some client file indexes.

Regards,

Bill Mason

0 Kudos
5 Replies
Highlighted
masonb
2 Iron

Re: Special group for index backups

Jump to solution

Johannes,

What you suggest is put in place on a lot of customer environments mainly for DR of the NetWorker server. Having one group linked to a separate pool and all clients linked to this group is common so that all indexes and a NetWorker server bootstrap so if DR is required the minimum amount of volumes can be read in to recover all indexes etc.

Using savegrp -O group name from cron or Windows scheduler is used as if you started the group within NetWorker you would attempt to backup all clients. The client creation procedure needs to be amended to ensure future clients are also added to the Index group so all indexes are secured during the run.

Regards,

Bill Mason

joka2
2 Iron

Re: Special group for index backups

Jump to solution

But is it ok to have this setup as a design of the running environment? Not just for DR purposes?

So my suggestion is not to have the backup groups initiate the index backup, but only rely on the index backup group.

So is that a common practice?

Thanks,

Johannes

0 Kudos
masonb
2 Iron

Re: Special group for index backups

Jump to solution

Johannes,

Yes it is fine for Production environment also, I just mentioned DR as an aside to show people are going what you suggest. It will just limit the volumes you would have to select from for any days index recovery. Depending upon your total index size it could take some time to fill a volume to capacity so this sometimes introduces a single point of failure to the environment but cloning these backups can alleviate this.

Whether DR or production the important thing is they are secured and we can recover quickly - this approach should satisfy whether recovering all or some client file indexes.

Regards,

Bill Mason

0 Kudos
joka2
2 Iron

Re: Special group for index backups

Jump to solution

Hi Bill.

And thank you for a clear and informative answer.

The point you mention about that it might take long to fill up tape is something that needs to be taken into account.

Regards,

Johannes

0 Kudos
ble1
5 Iridium

Re: Special group for index backups

Jump to solution

I use this approach in production.  We have around 400 groups which run on daily basis (and there is around 70 archive log groups running mostly every 1 hour) so if we would save index all the time we would have mess or simply had to split server into few of them.  In essence, on group properties you select option called "no index save" and group will not trigger index+bootstrap backup.  When running group with server client inside bootstrap backup will happen always.  If this group is disabled, it will override no index save setting on other groups.  index+bootstrap backup can then be either configured from cron by calling savegrp -O option or from GUI in recent (newer) version of NetWorker.

I have 5 index groups, each hosting different clients (limited to 130 clients each) plus group where I backup server, storage nodes and NMC host thus I backup indices once per day and bootstrap 6x times during the day.  The reason we mainly did this was not much due to index backups (we have VTL plus server is dedicated for itself thus no big deal with that), but due to a load towards nsrmmdbd which comes from a fact that after each group completes NW will run mminfo -B to generate bootstrap log.