Corrupt Index:What To Do?

[SOLARFLARE]

When I run nsrck I get the following about one of my indexes. Any ideas about what to do next.

missing file sr_i0
missing file sr_t0
nsrck: Check failed for client "name" (bad database header)

Thx.

 

[Datensicherung]

Hello,

use nsrck -F clientname

is this not successfully, you must recover the index.

dirk

 

[lesb]

I have encountered this on a couple of occasions where recover did not solve problem. Had to go to the index (in my case it would be /nsr/index/clientname) and
do "rm -fr ./db" to remove the index for that client.
When that was complete, do "nsrck -c clientname" and create a clean, but empty index. NOTE that the "-c" is lower case.
Les lesb - SUN SysAdmin

 

[bigkatt]

It has been recommended that you run nsrck -F clientname at least 3 times. Why three times, I don't know. I have run into a couple cases where the index has been fixed after the second or third time. I have also run into cases where it does not matter how many times you run the nsrck command, you will have to remove the client index and recreate it.
It is also important that you run the nsrck -F on your servers index. You must make sure that you turn off all groups for that period of time. You do not want the completed groups to try to write to you servers index while nsrck is checking the index. Our backup server usually takes 8 hours to complete the nsrck.

Good luck,
Rob

 

[wilsona]

Before you do anything else when you get this message, try simply 'touch sr_i0' to create an empty file for one it complains about. I have got this to work on more than one occasion, usually after running out of space for some reason. The index check, or whatever failed, then recovers and continues. Worth a try...

 

[brsluss]

how often should you run nsrck -F We are having a problem where we have a bcv copy of mail saved locally and when we back up this file system it takes upwards of 18 hours for 65 gig of data is this normal?

also has anybody ever tried to redirect a client index out of one client to another? we use the following scenerio

client save set

backup1 /opt

backup1 /mail

backup1 /mailserver

mailserver none

what we want to do to not have the index grow and get corrupt we want to redirect the client index backup1 with the saveset /mailserver to the mailserver client.

we tried this two times using the command save -c mailserver un the backup command line. and for our mailbackups it failed but for some reason it succeds for the nfs enviroment that we have!!! the failure stems by that on mail it does a full every day.