There is 1GB of physical memory available and 2 GB of swap available. I
would assume that one would see things exiting abnormally in the logs
though if the memory was bottoming out? In any case, the remote machines
seem to get backed up fine but when it comes to backing itself up then
there is a problem which makes me think that if it was a memory issue
then all of the backups would fail, not just the localhost... right?
You prompted me to go back and look at the logs, in doing so it seems as
if I get 2 X of each log per run which I didn't notice before and now I
am seeing some problems. I can't tell if this is a permissions problems
or if this is something else. Attached are the logs containing the
problems. A little more info is that as soon as I had these problems I
went and switched the name from localhost to the FQDN of the machine in
both .amandahosts as well as disklist in hopes this would fix the
problem. I hope I didn't make anything worse, it sure didn't make
anything better. 


On Thu, 2004-04-29 at 01:06, Paul Bijnens wrote:
> Kris Vassallo wrote:
> 
> > - I am getting debug logs in /tmp/amanda
> 
> What kind of debug logs?  There should be a log from amandad with
> the protocol packets in it, and sendsize log, and a sendbackup log,
> and to run{tar,dump} logs.
> Which are there, compare it carefully with one that did succeed.
> Maybe send it to the list, (or to me if you believe too much sensitive
> information is in it).
> 
> One of the stupid things that bit me once was running out of swap
> space: amcheck did succeed, but anything using more memory crashed
> without traces in the logs.
> 

amandad.20040428235901000.debug
Description: Text document

amtrmidx.20040429002907.debug
Description: Text document