No output error

[hockmarc]

Hi there!

I'm running Legato 6.1.2 on Solaris and and I'm trying to back up an Informix client. Everytime I start a backup, I get a "no output error":


08/16/02 14:45:33 nsrd: savegroup info: starting Test_DB (with 1 client(s))
08/16/02 14:46:19 nsrd: client.domain.de:INFORMIX:/onl_server_tcp saving to pool 'Pooldb' (DKM521)
08/16/02 14:46:22 nsrd: client.domain.de:INFORMIX:/onl_server_tcp done saving to pool 'Pooldb' (DKM521) 2977 KB
08/16/02 14:46:23 nsrd: client.domain.de:INFORMIX:/onl_server_tcp saving to pool 'Pooldb' (DKM521)
08/16/02 14:46:25 nsrd: client.domain.de:INFORMIX:/onl_server_tcp done saving to pool 'Pooldb' (DKM521) 62 KB
08/16/02 14:46:25 nsrd: client.domain.de:INFORMIX:/onl_server_tcp saving to pool 'Pooldb' (DKM521)
08/16/02 14:46:25 nsrd: client.domain.de:INFORMIX:/onl_server_tcp saving to pool 'Pooldb' (DKM521)
08/16/02 14:46:25 nsrd: client.domain.de:INFORMIX:/onl_server_tcp saving to pool 'Pooldb' (DKM521)
08/16/02 14:46:29 nsrd: client.domain.de:INFORMIX:/onl_server_tcp done saving to pool 'Pooldb' (DKM521) 62 KB
08/16/02 14:46:29 nsrd: client.domain.de:INFORMIX:/onl_server_tcp done saving to pool 'Pooldb' (DKM521) 62 KB
08/16/02 14:46:29 nsrd: client.domain.de:INFORMIX:/onl_server_tcp done saving to pool 'Pooldb' (DKM521) 2729 KB
* client.domain.de:INFORMIX:/onl_server_tcp ! no output
08/16/02 14:46:31 nsrd: shakespear:index:client.domain.de saving to pool 'Pooldb' (DKM521)
08/16/02 14:46:38 nsrd: shakespear:index:client.domain.de done saving to pool 'Pooldb' (DKM521) 120 MB
08/16/02 14:46:38 nsrd: savegroup alert: Test_DB completed, 1 client(s) (client.domain.de Failed)
08/16/02 14:46:39 nsrd: runq: NSR group Test_DB exited with return code 1.
08/16/02 14:47:10 nsrd: write completion notice: Writing to volume DKM521 complete

The networker client & the Informix module have been reinstalled on the client, so the problem is probably on client-side. Does anyone have an idea?

TIA,
MH

 

[RichS24]

you need to look along the route of this being a network issue. No output usually means that the connection between the server and client has been dropped at some stage, for whatever reason.
Check that duplex settings on the nic's at both ends are set to full duplex, for a start.

Try a manual save, and see if the same happens as well.

 

[keefb]

Check your card settings and network. Normally means connectivity issues. Regards
KeefB

 

[notthehoople]

Hi,

Just in case you are still having the problem....

"No Output" can also be produced if you don't have enough permissions to back something up. Perhaps you have a part of your database which is more secure than other parts, or the user who is running the backups has no access to the transaction logs...

I'd suggest:

- looking in /tmp/bar_act.log which should have some more interesting error messages in it and may show you where the problem is.
- if not, try backing up each of your blobspaces in turn and see which one is giving you a problem. You can specify a single blobspace by putting INFORMIX:/dbname/blobspace_name in the saveset of your NetWorker client.

 

[eskolnik]

I am working on an issue like this with Legato support, The work around was to change the client parellelism to 1. When I did this the backup works. Also note if the client is a windows client the drive must have SYSTEM with Full control.

 

[eskolnik]

I had the same error message while backing up a NT client. The saveset would retry X number of times and then fail.

We went back and forth with Legato support, the issue is still open, but changing the client parallelism from 4 to 1 stopped it from happening. We then installed the Networker Server 6.1.2 Aug 2002 CD and on the client, Changed parallelism on the client back to 4 and Wha-la NO problems.

Hope this Helps
Ed Skolnik

 

[edskolnik]

FR: isivapal

3026233 Error: nsrexecd: failed to write nulhandshake on 240 error code 154 LGTpa43607

ear Ed:


>>>>>>>>>>>>>
Indy,
What is the status of the binaries for debuging?
>>>>>>>>>>>>

Engineering is still analysing some issues and had given us their recommendations but we have further questions on some of their recommendation. In fact I have a conference call with the engineer on this in 20 mins.

Here is part of the dialogue from the engineer:

o The * <client>:<filesystem> ! no output messages appear
to originate from savegrp.exe when it attempts to read the
log created by nsrexec while executing the client save. The
message can be generated when either the log is zero length
or when no saveset completion message is found in the log
(successful to not). The &quot;no output&quot; message is documented
in savegrp(1m).

o The client log files (nsr\tmp\sg.<group>.<client>.XXXXXX)
are normally deleted but can be retained if the debug level
is greater than 1 (eg. -D2). This would require running
savegrp manually.

I will let you know once we have clarified a few more points with the engineer. Thanks Indy

 

[eskolnik]

Legato has come up with a HOT fix that transmits keep alive data on the control ports so the firewall does not close the port before the same complete's. We installed it on the backup server (Solaris and on two Windows 2000 clients . It works for us

Ed