Bacula-users

Re: [Bacula-users] Solaris "Packet size too big" failures

2008-12-03 13:47:25
Subject: Re: [Bacula-users] Solaris "Packet size too big" failures
From: Jason Dixon <jdixon AT omniti DOT com>
To: Allan Black <Allan.Black AT btconnect DOT com>
Date: Wed, 3 Dec 2008 13:44:47 -0500
On Wed, Dec 03, 2008 at 06:35:51PM +0000, Allan Black wrote:
> Jason Dixon wrote:
> > On Thu, Nov 27, 2008 at 01:01:12AM +0100, Kjetil Torgrim Homme wrote:
> >> Jason Dixon <jdixon AT omniti DOT com> writes:
> >>> On Wed, Nov 26, 2008 at 11:59:41AM +0100, Kshatriya wrote:
> >>>>> recently.  Since then, we've encountered frequent failures of the
> >>>>> catalog backup job which reads from the local FD.  They always
> >>>>> fail with a "Packet size too big" error, which seems to be more
> >>>> Check if you are using a different MTU size than the rest of your
> >>>> hosts in the network.
> >>> This is only for the localhost FD.  Traffic should never hit a
> >>> physical interface.
> >> I don't know if it is relevant, but the MTU for lo0 on Solaris is 8232
> >> I wouldn't use "localhost" anyway -- I prefer to use globally unique
> > Good idea, I've put this into testing.  I'll report back our results in
> > a few days.
> 
> I don't think this is the problem - 'Packet size too big from XYZ' is a
> Bacula error, caused by the SD rejecting a message from the FD. Data from
> the FD to SD is encoded in 2 packets; a 'header', which contains (among
> other things) the size of the second packet, which contains the data.
> The error is caused by the size being out of range (max is 1,000,000).
> 
> Earlier in the backup there was a read error from the FD (No data
> available). The ENODATA is generated internally by Bacula when it reads 0
> bytes (a read returns <= 0, and errno == 0). This would suggest to me that
> there is some sort of problem with the FD.
> 
> Is it possible that the generation of the catalog backup is taking so long
> that the connection between the FD and SD is timing out?

Wouldn't it fail consistently in that case?  It hasn't failed the last
two nights.

-- 
Jason Dixon
OmniTI Computer Consulting, Inc.
jdixon AT omniti DOT com
443.325.1357 x.241

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users