Networker

Re: [Networker] Savegroup Parallelism problems

2003-07-31 11:28:46
Subject: Re: [Networker] Savegroup Parallelism problems
From: Carl Farnsworth <cfarnsworth AT SYMPATICO DOT CA>
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Thu, 31 Jul 2003 11:28:34 -0400
I think this is not really a problem with Savegroup or Client Parallelism -
do you also see time-out errors in the Savegroup Completion report?  If you
were to look at the daemon.log, I think you'll find there are time-out
messages in there, then it may restart the same session, or go on to the
next one.  The thing is, there is really only one session saving at a time -
 it's either a bug with the GUI, or an internal auditing bug where the
saveset gets abandoned, but somehow it's not cleared completely.  If you
watch the monitor window closely you'll see the size of one stop growing,
then a new session starts up.  Do these eventually go away some hours after
aborting the group?

I saw the same situation not too long ago, but in a Win2K environment, all
6.1.3.  In this case the client is behind a firewall, and I'm assuming it's
network connectivity which causes this problem - especially since when
first installed, the client behaved perfectly!!  Then one day, it simply
stopped working - somebody changed something, somewhere ...  Still no
answer, but double-check network config and routing between client and
host - do either have multiple NICS?

Carl Farnsworth
cfarnsworth AT digidyne DOT ca


On Wed, 30 Jul 2003 09:26:15 +0200, Frank Altpeter <f.altpeter@BROADNET-
MEDIASCAPE.DE> wrote:

>[networker-6.1.3 on linux-2.4.21]
>
>hi there,
>
>There seems to be a problem with the control of the Savegroup
>Parallelism values at my site, which makes the backup of some
>machines almost impossible...
>
>I have defined a group, which has savegroup parallelism set to '1',
>because the host is too slow to handle more than one save process at
>the time.
>The same value is defined for this host in its client definition.
>
>But, if i start the group (which runs for many hours), i get
>sometime the effect that there are 3 partitions saved in parallel,
>which shouldn't happen with savegroup parallelism set to 1.
>
>And, the best thing... one of the partitions is saved twice, the
>same time...
>
>
>Looks like this:
>
> host.domain.net:/disk2 saving to pool 'Filesystem Backup' (r2d2.fs) 3154
MB
> host.domain.net:/disk2 saving to pool 'Filesystem Backup' (r2d2.fs) 1283
MB
> host.domain.net:/disk3/wwwstat saving to pool 'Filesystem Backup'
(r2d2.fs) 899 MB
> host.domain.net:/disk3/www saving to pool 'Filesystem Backup' (r2d2.fs)
498 MB
>
>
>And... after manually stopping the running group, i get the 'group
>aborted, 1 client(s) failed' message, but the above saves are still
>running (and saving)...
>
>
>The client system is a Solaris 8.
>
>Any hints appreciated...
>
>
>With kind regards,
>
>        Frank Altpeter - 2nd Level Support

--
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.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

<Prev in Thread] Current Thread [Next in Thread>