On Tuesday 25 July 2006 10:44, Gene Heskett wrote:
>On Monday 24 July 2006 16:00, Kevin Till wrote:
>>Gene Heskett wrote:
>>> Oh oh folks:
>>>
>>> I put 2.5.1b1-20060723 in on the server (this machine) yesterday,
>>> got a near total failure last night.
>>
>>Hi Gene,
>>
>>o when you run amcheck, did it pass?
>>
>>o did the amandad on the client respond at all? Check if there
>> /tmp/amanda/amandad*.debug file created?
>>
>>o If so, any error in that file?
>>
>>o are you using bsd (the default) as the auth?
>>
>>o Try to add "server_args             = -auth=bsd amdump" to the
>> xinetd/amanda defintion. More detail info on docs/howto-auth.txt
>>
>>> ----------  Forwarded Message  ----------
>
>[snipped due to formatting damages from an over exhuberant line
> wrapper someplace in the message handling chain]
>
>Is this thread to be allowed to die?  It was a very real problem for
> me. I can reinstall this latest snapshot to verify its broken-ness I
> suppose...  Cleanup seems to be minimal, kill all the gtars on the
> client & reinstall 2.5.0-20060424, but it sure mucks up the
> scheduling for a few days.  FWIW, the client is still running
> 2.5.1b1-20060722, and worked just fine last night.  This is the one
> with the memory leak? No, thats the one that fixed it.
>
>So it blows up if the server is 2.5.1b1-20060723, and the client is
>2.5.1b1-20060722 in this case.
>
>>>From the 723 ChangeLog:
>2006-07-23  Jean-Louis Martineau <martineau AT zmanda DOT com>
>        * client-src/sendbackup-gnutar.c: Fix bug found by splint.
>        * client-src/sendsize.c: Fix bug found by splint.
>
>So what did this break?  It broke both machines, although it only left
> a gtar per DLE running on the client box, but not on the server box. 
> And it was the server box that was running 0723.
>
>Me puzzled...  The canary died.

I ran 2.5.0-20060424 on both boxes for the next nights run, and it 
seemed to run normally but trying to play catchup, which was expected.
So last night, having some blood to spare, I ran 2.5.1b1-20060725 on 
both machines.  That also seemed to be ok, but still playing catchup, 
using 100.1% of the 'tape' and the report indicating 18 level 0's had 
been delayed.  A 'balance' report says it will overflow 13% again 
tomorrow night, so its much closer than after the fiasco 2 nights ago 
when it was saying +78% for the next run.

So this new canary (amanda-2.5.1b1-20060725) seems to be quite healthy 
so far.

-- 
Cheers, Gene
People having trouble with vz bouncing email to me should add the word
'online' between the 'verizon', and the dot which bypasses vz's
stupid bounce rules.  I do use spamassassin too. :-)
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2006 by Maurice Eugene Heskett, all rights reserved.