Force index to save with data

check all you media pool setup and try to see if there is a pool whose properties include "Pool Type=backup" and "Save Sets=bootstrap index".

delete the pool.

I am afraid we cannot guarantee that the data and index are saved to the same tape, but we can let them go to the same pool. Please understand if there are more than one appendable tapes in the destination pool, it will pick up tapes randomly. That is to say it is possible for the data to go to one tape and index to go to another tape.

If you are trying to save them  into the same pool, try Kevin’s suggestions. If you still want to save them to the same tape, one thing you can do is to make sure only one appendable tape is mounted in the destination pool when backing up the data. However, this is almost impracticable.

DaveK,

You can set the index backups not to save within the groups by selecting the option “No Index Save”.

You can then create a group, link all clients to it and select “Index Only” option and start it at a time when no other backups are running.

This way you should get all index backups and a bootstrap for the NetWorker server on the same tape. A lot of people choose something like this for DR purposes – and run after/before backup window so you catch any restarts on that day.

A word of caution though is depending upon the size of your environment and if you use a specific pool linked to this Index group you can end up with a lot if indexes/bootstraps on the same volume before it fills to 100% becoming a single point of failure so with smaller environments is usually best to pick a pool which is the most used or you can swap out the tapes on a regular basis.

Regards,

Bill Mason

DaveK

If you have index savesets going to Default pool and you don’t use this as a target for backups then there is something wrong in your configuration (maybe you have linked pools to specific drives and you don’t have one for the NetWorker server itself linked).

No Index means we omit the saving if the client indexes as part of the backup so just save the data. As I described you can then set up a specific Index save so you can recover client indexes when required – it just means you will have one or two backup whereas if you backup the same client multiple times in 24 hrs you could end up with many more than 2.

Archive services is different to backup where basically the backups never expire and are not common within environments.

I believe the ”management node” is the NetWorker server itself and it should be secured but the main thing you need for recovery of this server in DR situation is the bootstrap information and then the Indexes. It’s very important you know the bootstrap backup details (which volume, pool, ssid, file number) where the backup is located. Having it in a group alone or with others is not an issue - as long as it’s in an active group – you could also use this group for your index backups using save index only option so you don’t create single point of failure for a tape/volume.

Regards,

Bill Mason

I don't have any save sets specified in any pool. All pools I use are set to 'backup', none are archive. Any save sets are specified in a particular client.

It is OK for indexes to go to the same pool, but not to different pools, such as 'default'. At least the tape would be in the drive and I would not see swapping just for index storage.

If I use the "no index' option on a pool, does that just mean that no index gets saved on a tape, but I do get an index saved on the management node? It is rare, but I do have a restore once in a while. Most of the index browse settings are at least one year, some are seven. I would like to have the ability to browse at least a year so I know which tapes I need to get back from off-site storage.

What does the 'archive services' option on the client mean? To me, an archive and a backup mean the same thing.

Should I back up the management node to its own pool? Right now it is set to the same pool as the daily production app servers.

I completely agree with Mr. Manson, its go to direct all the index data on a single tape and also for saftey purpose do keep a clone copy of that tape. So whenever there is a index restore required automatically the required tape will be used.

As requested by you, We cannot gaurantee that index data will be saved in the same tape(that highly depends on space available on tape) but you can restrict index Backup to be saved on the same pool as your data does.

I should have specified I wanted to save to the same pool, not the same tape.

I don't see any option for "index only" in the client or the group settings. I just looked at all my clients. Is this a directive or some additional parameter setting?

I did have several clients with "archive services" checked. I unchecked those.

In the pool configuration, I do have 'store index entries' checked on all of the items.

Group Properties -> Advanced -> Options. See below:

Thank you.

Here is a related question. I have moved a NDMP client to another group from its original group. However, when the backup runs, it stores the data on the correct tape, but the index information goes to the original group's tape, forcing another tape swap. So, rather than calling for the default tape or a bootstrap tape, it goes to a group I no longer want the index to be saved with. Can I control that? Also, if I move data to another folder and a new group, it wants to make a full backup because it has not seen this particular arrangement before. I already have a full backup under a different name. How can I stop this wasteful backing up? It is on the order of 4-5 TB of data that I don't need another full backup of.

You control that via pool properties - which you or someone else did in the past, but it was not done when you moved the client to the new group.