Recurring problem on index space.

Question

friends,

I have my backup server running on RHEL 4 and networker version 7.3.2 only.. and my index location is /nsr/index and mm on /nsr/mm path

now, i am having a serious issue on space growth on /nsr/index path, because of my requirement of high retention policy.

some folks adviced me to do "expiry of indexes". Could someone detail me on index and this operation also.

Thanks

Biku1 · Accepted Answer

Hi Vijay,I want to explain you that what are these policies:1)Browse Policy :It is the time period upto which the server have the indexes for the client and when we want to restore the data for the client then we can browse the Data upto the file level in this period of time.Retention Policy: This also a time period but it is always greater than the browse policy or equal to the browse policy but not lesser than the browse policy.In this Retention policy time period we can not be able to browse the data upto file level.but we can browse the data upto the Savset level that we define in the savset of the client resource.But in this also we can restore the each file by mentioning the file path that we need to restore.So in this policy also we can restore the data safely.For example:Browse Policy Retention policy Data Type for recovery30 Days 60 Days browsable= 30 Days,Recoverable 31-60days30Days 30 Days Browasble=30days,Recoverable upto 30 days2) If the data is out of retention policy then also you can do the restore but on one condition that the tape should not relabled.It will show you the data in the tape as recylable but then in that condition you need to run the scanner command.scanner -m orscanner -i3)About the indexes space: You can change the indexes path in the client resource.And when the next backup run it will save the indexes of the client on the new location that you mentioned the index path of the client resource...Hope all this will Help you

Hari5 · Answer

Expiry of indices can be done by removing the oldest cycle of individual save sets from the NMC gui. You need to go to Media>Index to do this. This is very time consuming and is a stop gap solution. On the long run you have two options either change your browse policy for your backups to a shorter time or relocate your indices to a file system with more space.

Relocation of indices can be done by copying the index directory (this would be /nsr/index/ ) for the individual clients to a new file system. The client configuration for the client being moved would have to be changed. The index path in the globals tab would need to be updated with the new path of the index. Please ensure that there are no backups going on when you attempt this. Alternately, you could move the /nsr/index/ directory to a new location and create a soft link to the new location(/nsr/index/ -> /newnsr/index/ ). Ensure that networker services are shut down when you do this.

dpinink_silva · Answer

And also, you could lower your browse policy. Backups will still be avaliable on media database, but will not be available on index. So it's a matter of balacing space needs with the need of having index available for quick restores.

Pipeliner1 · Answer

I agree, set your browsable to something like 6 months and then keep your retention to 10 years or whatever the need is. I have recently had to to this to manage an index size explosion due to a legal mandate. I set it individually on each client using the nsradmin utility. It's a 5 minute job:

type nsradmin and at the nsradmin prompt type the following:
. type:nsr client
show name; browse policy
print

This will give you a listing of all of your client resources and their current browse policy. Then type the following to update:
update browse policy:NAME_OF_POLICY_HERE

and then simply hit y and enter for all of the resources to change. For me, I am comfortable doing save set recovers so I kept the browse policy to 6 months to save on the size of the client file indexes. I just went into the policies part of the NMC and defined a 6 month policy and the name of that is what you use in the update command. Hope this helps.

vijayscsa · Answer

Thanks for all your inputs. I have 3 different retention policy for daily(year),weekly(3years) and monthly(10years). For all, shall i go ahead with 6 months as Browse policy ?

also, i have confusion how to retrieve the data from tapes, if the required data is out of retention policy.

Thanks Pipeliner

vijayscsa · Answer

Hi BikuThanks for your inputs.But,my only version is 7.3.2 and only OS is Linux for NetWorker server. No storage node configured. As, i was forced to keep the backup completed data tapes for 10 years, with out recycling,,,, i can use scanner command if the data requested is only in the tape and not in the index.....Kindly let me know, which is the GUI based restoration utility in linux on 7.3.2 for restoring the datas. also, what is the term SSID on networker.Thanks in Advance

dpinink_silva · Answer

GUI based restoration for linux/unix: nwrecoverSSID: the unique indentifier of a saveset thas has been backed up.If you need more info about these questions, I would suggest to open a new topic

NetWorker

Recurring problem on index space.

Was this post helpful?