nsrd will not stay running after server crash

[cstorm]

One of mey Legato Networker servers was hung and had to be rebooted. Networker nsrd process dies after 5 seconds. I am unable to issue any Networker commands. Is my only alternative to reinstall Networker software?

 

[605]

You do not need to do this. The only problem that will prevent
nsrd from starting up is a defective media database. This
should be noticed in the daemon.log file.

To repair this, you must run the disaster recovery procedure
(mmrecov) after you have deleted the media database. Then
recover the client file index backups with "nsrck -L7". The
manuals will explain you more details.

If you have done this already, you will notice that this is
very easy. Good luck!

 

[TDun]

This is the action if it is u/s.
However, if you look in the daemon.log it will hopefully have a message telling you what problem nsrd had.
This may be something that is more easily fixed.

Myself, I would guess that nssr.res has been corrupted/truncated in the crash. If this is so, you will need to follow 605's advice.

 

[605]

Some years ago i have done a lot of tests to find out whether
a corrupt resource file will prevent nsrd from getting started.
I never could create such scenario, but i do realize that such
could of course be possible.

Do not forget:
- Even missing resource files will not matter. Remember,
that for the first initialization, no such file will
exist.
- In NW 7.x, a corrupt resource file will be isolated. This
of course will be logged.

 

[birky]

Hi

We had this problem last year and it did indeed turn out to be a corrupted media db as 605 says.

 

[cstorm]

I renamed the /nsr/mm directory and issued the scanner -B command to get the latest bootstrap saveset id and file number from the DLT drive/tape. However, mmrecov does not see my tape drive. It only sees the internal drive (4mm). I tried adding my tape unit via jbconfig, but it causes the server to reboot. Any suggestions would be appreciated.

 

[605]

Make sure that inquire will see all devices on the NetWorker
server.
Those devices/jukeboxes must be be configured before you run
mmrecov. However, with jbconfig, you cannot install a stand-alone
tape drive. I wonder why jbconfig will cause a reboot, even if
it will be used to install a stand-alone device, but i never
tried this.

It could also be that a binary has been corrupted. In such case,
you of course need to re-install the software.

 

[cstorm]

Thanks for the help. I got lucky on this one. I had forgotten that as a precaution I have been backing this Networker server up using NetBackup. I was able to restore /nsr and then use the mmrecov procedure to get things back in order. Thanks again.