Help Needed: NMM Client unable to restore browsable savesets

Have you tried recovering a later index and had the same problem?

That doesn't necessarily follow, your savesets will be recoverable if they are within the retention policy but they will not be browsable until you restore the indexes if the backups are older than the browse policy, which you give to be 3 months.  From that figure anything older than 18/10/10 would need to have an index recovery, assuming your retention is longer than your browse period, so I assume what you meant was "Savesets from Mid-Oct to Dec were alright (browsable) and were able to be recovered".

If you have backups prior to this you can either restore indexes as you have been trying to, or you can do a saveset recovery (which does not use the indexes).  I am assuming from what you have been saying that you have non-browsable backups which you are trying to recover.

I would suggest you take a look at the index backups and try to restore the next full index backup after the date you are trying to restore and see if you are able to recover that - the later full index backup will have details of previous backups, not just relate to that particular time; use:

mminfo -q "name=index:clientname,level=full" -ot

The user wanted to restore savesets from April to December. Savesets from Dec to Mid-Oct were alright, the rest can't be restore...

Unfortunately in my environment, by standards browsable policy have to set at 3months, any savesets more than 3months old can't be restore by the client.

Hi David,

These tapes are set for few years of retention...

Actually I already identified the indexes that I wanted to recover. But just don't know why after few scanning on the tape, i'm still unable to recover the indexes..

Last night i loaded the tape into the tape drive with the indexes that I wanted to restore...I executed this command

scanner -i -vvv device_name

it took about 10hours to complete..

As according to the support, after that I should be able to recover the indexes within the command

nsrck -L7 -t"mm/dd/yy" client_name

it came out to be same problem again...index for the client can't be recovered. Hours of efforts are in vain again...@.@

I realized the scanner command is very time comsuming...is this scanner then nsrck -L7 the only solution we have to change recoverable savesets to browsable? do you know of any alternative solutions that can do the same job?

TahWee

The correct way of recovering indexes is to use nsrck -L7 as you have been doing.  Scanner -i is the old way of getting indexes back which reads through the tape and builds an index from what it finds there (note that you would run scanner -i on the tape which has your data on it and not the one that has your indexes on it) - scanner is a time consuming way of recovering your indexes and should only be used as a last resort.

If you want to recover indexes to browse data bear this in mind; the index data at a point in time does not disappear after you back it up but will stick around until the saveset it relates to passes outside its browse period.  Therefore if you wish to recover your indexes you don't have to recover the nearest point in time, you can recover later indexes (in your case any index backup from the next 3 months) and ideally it is better to pick a full index backup to recover as this will minimize the number of index savesets you need to recover to get your data back.

If you have tried recovering several different index backups and have experienced issues with each one I would suggest the problem is something more fundamental.  If you can add -v to the nsrck command then post the output here we may be able to advise further.

I see...

what is this -v for in the nsrck command? i can't seem to find the use of the -v from the command reference...

this will not affect the historical data? because this is a production system...

With Networker commands -v commonly means give verbose output.

esg112266 had really saved my life..

Scanner is something you should almost never have to use these days.  Usually nsrck -L7 -t time clientname will recover the indexes which is the correct way of recovering indexes these days (since version 7.1?)

yes, i understand that.

I was troubleshooting with the PREM, and the above solutions was the best so far..

as i mentioned, when i tried to execute the nsrck -L7 -t time client_name, it kept giving me error messages...then i discovered esg112266 that matched my conditions..

I try to upload the nsrck -v, when the server is available, we then investigate again?

Anyway, David really appreciate your continuous advices =)