On Fri, Sep 20, 2019 at 13:28:42 +0000, Robert Reilly wrote:
>
> First one is the failed backup second is the successful backup
>
> https://gist.github.com/rreilly-edr/debff9f0ce3a1759993083b54a966cda
It would still be helpeful to see your actual disklist file, but it
looks like you have two Amanda clients configured via SSH and one using
BSDTCP auth.
In the successful run, all three clients respond to the "estimate
server" request essentially immediately and the planner shuts down
cleanly, around 2 seconds after it starts.
In the failed run, the BSDTCP client reponds with its size estimates
within about 18 seconds, but the other two clients never send any
calcsize data back to the planner, and after (I believe) the etimeout
period expires the planner gives up and shuts down.
So, I guess the next thing to look at is the exact status of the
"calcsize" processes on the clients during this waiting period. Data
from those processes has to go back to the server via a complicated
chain (simplified, this chain seems to be something like: calcsize ->
sendsize -> amandad -> sshd ---[network to server]--> ssh-auth-driver ->
planner), so it could be somewhere in there....
Or it could be that "calcsize" just isn't running/terminating, for some
reason (though this seems less likely, since the problem seems to have
something to do with upgrading to 3.5.1 on the server side...).
Anyway, in addition to looking at processes on the client and
strace/lsof of the calcsize (and sendsize) processes during the waiting
period, more generally looking at the "sendsize" debug files on the
clients might tell us have far along the clients are getting and if
there are any errors along the way.
Nathan
----------------------------------------------------------------------------
Nathan Stratton Treadway - natha...@ontko.com - Mid-Atlantic region Ray
Ontko & Co. - Software consulting services - http://www.ontko.com/
GPG Key: http://www.ontko.com/~nathanst/gpg_key.txt ID: 1023D/ECFB6239
Key fingerprint = 6AD8 485E 20B9 5C71 231C 0C32 15F3 ADCD ECFB 6239
