Errors when starting nwadmin.

[Guest_imported]

When I try to start nwadmin/networker, the following errors appear at the daemon.log file and the program starts with a server not found message.<br><br>nsrd: server notice: started<br>nsrmmdbd: error adding btrees to ss (an invalid slot number)<br>nsrmmdbd: WISS error, an invalid slot number<br>nsrd: unable to start nsrmmdbd<br>nsrd: shutting down<br>nsrd: successful shutdown<br><br>Further up in the log, the following messages appear:<br><br>nsrmmdbd: WISS error, database was not shutdown cleanly<br>nsrmmdbd: media db must be scavenged<br>.<br>.<br>nsrmmdb: media db scavenge successful<br>.<br>.<br><br>Any help with this issue will be greatly appreciated.<br><br>Antonio Trejo.<br>

 

[avedis]

If you are using UNIX, this shoudl work...
If your using NT, you can try, but I do not know...

When the server is down type at the UNIX command prompt:

nsrck -m

This will rebuild the media database, From the error you are getting, it looks as though you have media database corruption.

You may get a warning or error when you do the above command saying &quot;Media Database Is Not Responding&quot; or something of similar. This is normal since the server is down. The nsrck will still do it's thing.

Depending on how big you media DB is, it may take anywhere from 5 minutes to an hour.

If it goes over twho hours, you may kill it, and try again, if it hangs again, possibly bad damage to media database, and you may have to recover it from tape.

B.Wolf

 

[cfowler]

Tonny21,

Hi there. Guess what I'm doing this weekend? I have encountered the same problem that you have so gracefully described. As opposed to doing a partial disaster recovery, I make it policy to take a copy of the nsr.res file, that is currently working and proven, and preserve any time I make a change to the current config. What I have done, is restore the nsr.res over the existing one with the problem. This gets me all of my groups and clients back. I'm not running again. The tapes still have all the information on them that they had before. This was suggested to me by Support. I think that I would have to agree with avedis though, it's more thorough doing it through the mminfo command. However, if you don't have time, we have quite a few servers and our indexes run at aroung 10GB for a 4 month period, then this is quicker.

By the by, to avedis, that command does indeed work with NT. Thanks to Legato and the override bug in ver 5.7.0.0 build 112, I've had to use it once every month now. This is the first time I have not used that command.

Hope this helps,

CF