What about index corruption?

From Backup Central


Index corruption can be a fatal problem with Networker. Nsrd will abort prematurely if there are problems with the indices. Stopping networker, then running nsrck -F is the best first step. From reading the man pages on nsrck, mmrecov, & nsrim -- as well as previous email in the Networker archive -- other commands to try in fixing corrupt indices are, in the order of their power: nsrck -m mmrecov [From the bugs section of the man page on mmrecov: Mmrecov is mis-named, causing unsuspecting users to use it (and its brute force features) when it is not needed. A name like "recover_server_index_or_media_index_when_either_ is_missing" is more descriptive. Note that any part of the bootstrap save set contents are recoverable using normal recover procedures provided that the server's on-line index and media index are in good shape.] nsrim -X On a related issue, Stan Horowitz posted (20 March 2000): At the end of this process, the /legato file system was at 100% and the nsrd daemon stopped. After shutting down the remaining daemons and cleaning up the indexes, we attempted to restart Legato. The result was the following:

root 10708 57170 1 09:35:06 pts/8 0:00 grep nsr root 55320 1 0
 08:46:41 - 0:00 legato/bin/nsrexecd -s localhost -s
 r0011isp.st11.meijer.com

As one can see the nsrd daemon did not start. Attempts were made to start it manually, but also failed. Any ideas where I can look to find a solution and/or what needs to be done to get Legato going again. Thanks in advance for any suggestions. Your NSR server's media database is corrupt, probably because there is no disk space left for NSR to operate. Add more disk space for NSR. Note that NSR needs at least twice the available disk space as the largest client index you have. then try to running "nsrck -X" should fix the problem. Lynn Glessner, writing on 16 March 2000, shared that she is one of several who use nsrck -F as part of a regular maintenance plan on her Legato indices. I tend towards over-maintenance

Also see section What are "WISS" errors?