Full backups do not finish...end with "unexpectedly exited"

[LilOrley]

Folks,

We've recently had a problem crop up. We have full dumps scheduled to start on Sunday night, and incrementals the rest of the week. The full dumps start off O.K., but later seem to "hang". Eventually, an error gets put into the daemon log file:

savegrp: myserver.blah.gov:E:\ unexpectedly exited
savegrp: myserver.blah.gov:E:\ will retry X more time(s)

(where X is the number of retries left).

Several hours and retries later, the savesets eventually abort and the only other bit of info put into the messages log is:

savegrp: myserver.blah.gov:E:\ ! no output

We're running Compaq Proliant servers with NT 4 sp 6a and Networker 6.0.1.

Any ideas?

Tnx!

allen

 

[Giggle1]

Which platform was the client on? We are having the same problem with only certain clients.

 

[LilOrley]

All of our servers are Compaq ProLiants (Intel based) running NT 4.0. We have one of these servers dedicated as our backup server and the rest (7 in all) are clients. We are running Networker 6.0.1.

Interestingly enough, we had successful full backups last weekend. Hopefully, this mysterious problem of "unexpected exits" has unexpectedly gone away.

Someone else at our facility has commented that Networker seems particularly sensitive to network anomalies, and we're running on ATM, which could be the problem. It's hard to say. We get no other information from the logs other than "unexpectedly exited".

Let me know if you find anything else out on this.

Tnx!

allen

 

[Giggle1]

Hi,

I found the tapes are almost full when the error occurred and the backup seems to work when it starts writing to new tapes. I found this to be the case sometimes with our clients which are NetApp filers and giving the same problem. But we dont get the problem when using a Unix back-up server running 5.5. The difference between the two seems to be something at the end. Uptil then, the same amount of data is being written to the tape. Also, if you run mminfo it will list those backups as complete with their SSIDs as well. For us at the moment Legato is looking at the problem. We do have a couple of NT server clients as well. We have tried only 1 of them so far, and I dont think it has failed.
Please tell us if it happens again. We have a switch between the filer and the server, which is also an NT 4 SP6 with Networker version 6.0.1.

thank you

 

[LilOrley]

Thank you for the info, and please let me know if Legato comes up with a solution for you.

Thanks!

allen