Networker

Re: [Networker] aborted due to inactivity

2004-01-06 14:44:02
Subject: Re: [Networker] aborted due to inactivity
From: robert.maiello AT THOMSON DOT COM
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Tue, 6 Jan 2004 14:43:57 -0500
Joel,
You could try a bigger inactivity timeout but..

It sure sounds like the tcp idle timeout value is being hit.   When backing
up through a firewall its often necessary to increase this value on the
firewall from 1 hour to say 4 to let large clients finish.   When Networker
does a save on a client its sets up several tcp/ip streams, some of which
are control sessions that remain quite idle as a save transverses large
filesystems/directories.  It is these control sessions that sometimes
timeout (and are needed for the client to say "hey I completed this saveset").

With no firewall here I'd guess to say maybe, as others said, that the
Windows tcp idle timeout value is being hit.    Unfortunately I'm not
familar with setting or what it is on Windows.  A search on "tcp idle"
yielded 1 message in the archives.  It referenced several links on Windows
tcp values:

http://support.microsoft.com/default.aspx?scid=KB;en-us;q120642

http://www.microsoft.com/technet/treeview/default.asp?url=/technet/itsolutions/network/deploy/depovg/tcpip2k.asp

http://listmail.temple.edu/scripts/wa.exe?A2=ind0207&L=networker&P=R23861&I=-3&X=122E6D166C225D1EC8


I'm sure other people here running Networker on Windows may have increased
these this value and are familar with it ??

Robert Maiello


On Tue, 6 Jan 2004 14:11:39 -0500, Joel Fisher <jfisher AT WFUBMC DOT EDU> 
wrote:

>Hey Robert,
>
>When the backup is started from the server it does fail a 42GB then
>restarts and fails at 42GB again.  No firewall involved.  I've kind of
>ruled out a network problem, because this is the only one of about 20
>large(100GB) volumes on the cluster that fails on a consistent basis.
>
>Still searching....
>
>Thanks,
>
>Joel
>
>When you say it always fails at the 42GB mark; does it always fail after
>the same amount of time?  If you have serveral retries set, do you see
>it
>back up 42Gb, then timeout, then back up 42GB again?
>
>If so, this would indicate an external severing of the control
>connection..
>
>First question from Legato will be "is there a firewall involved?"
>
>Robert Maiello
>Thomson Healthcare
>
>On Mon, 5 Jan 2004 14:17:35 -0500, Joel Fisher <jfisher AT WFUBMC DOT EDU>
>wrote:
>
>>Hey All,
>>
>>
>>
>>I'm wondering if anyone might be able to shed some light on a problem
>>I've be troubleshooting.
>>
>>
>>
>>Server: Sun E450/Solaris2.6/SBU 6.1.3
>>
>>Client: Compaq/Win 2K3 cluster/Legato 7.0 client
>>
>>
>>
>>I get the below failure on full backups(incremental backups complete
>>successfully).
>>
>>
>>
>>* client1.nt.wfubmc.edu:P: 1 retry attempted
>>
>>* client1.nt.wfubmc.edu:P: 01/02/04 08:53:23 nsrexec: Attempting a kill
>>on remote save
>>
>>* client1.nt.wfubmc.edu:P: 01/02/04 08:58:23 nsrexec: Attempting a kill
>>on remote save
>>
>>* client1.nt.wfubmc.edu:P: write: Broken pipe
>>
>>* client1.nt.wfubmc.edu:P: aborted due to inactivity
>>
>>
>>
>>I've increase the savegroup timeout to 360 and it made no difference.
>>
>>
>>
>>I can run the backup directly from client1 successfully, but when I
>>start it from the backup server it fails.  When it fails it always
>seems
>>to fail at 42GB.  So I thought maybe there as a bad file or something
>>like that causing it to hang, but when it ran successfully directly
>from
>>the client that killed that theory.  I'm kind of stumped right now.
>>
>>
>>
>>Anyone ever have a similar problem?
>>
>>
>>
>>Thanks,
>>
>>
>>
>>Joel
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>--
>>Note: To sign off this list, send a "signoff networker" command via
>email
>>to listserv AT listmail.temple DOT edu or visit the list's Web site at
>>http://listmail.temple.edu/archives/networker.html where you can
>>also view and post messages to the list.
>>=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
>
>--
>Note: To sign off this list, send a "signoff networker" command via
>email
>to listserv AT listmail.temple DOT edu or visit the list's Web site at
>http://listmail.temple.edu/archives/networker.html where you can
>also view and post messages to the list.
>=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
>
>--
>Note: To sign off this list, send a "signoff networker" command via email
>to listserv AT listmail.temple DOT edu or visit the list's Web site at
>http://listmail.temple.edu/archives/networker.html where you can
>also view and post messages to the list.
>=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=