Help, I have a client that keeps failing backups

Do you also see and NSR peer information OR authentication not supported kind errors? You need to set to oldauth first to get rid of those.

For this, the problem is that during the backup the connection is getting reset or broken. You can set parameters like NSR_KEEPALIVE_WAIT & timeout activity to make sure the server waits a bit longer for the backups to happen.

Might even be some interruption of the RPC communication between the server and client.

what can cause that?

Most often firewalls. There are parameters on how long RPC connections are kept alive, regardless of if they are active or not, on some firewalls.

HI,

If I understand correctly, the client is also the NetWorker server and there is a database running on this same host. If this is a busy host, it could be running out of system resources and a TCP/IP connection to itself is getting closed.

Now, that being said, if it is always the same directory, it is more likely something about that directory that is causing the failure.

I suggest running a manual save of the failing directory on the client in verbose and debug mode to troubleshoot this. There may be a problem with a file in that directory. This way you can see where in the filesystem the failure occurs. If debug level 3 does not reveal enough detail, you can go up to -D9, but this produces a huge amount of debug data.

Eg:
save -v -D3 -s networker_server -b pool > save.log 2>&1

where:
networker_server is the NetWorker server hostname.
pool is a valid backup pool for this client.

Check the end of the save.log file to determine the last file and directory that were backed up. Check the files in the directory. Try copying them to see if a bad file can be located.

If there is a problem with the filesystem, an OS filesystem check may be needed.

Hi ,

Fianlyy did you found a fix for this problem ? Because we too have a similar issue in backing up .BAK files in our environment.

Thanks

Hello - I have a similiar issue on some clients since we changed our storage node set up.

Only subset of windows hosts affected - all solaris hosts OK.

Where is NSR_KEEPALIVE_WAIT actually set? On the server or on the client also, storage node also?

Thanks
Nick

On the client.