Stefan G. Weichinger a écrit :

Stefan G. Weichinger schrieb:

A DLE with ~180GB fails to be retried on the second tape (LTO-3-changer) with

FAILED [data write: Connection reset by peer]

I already decreased tcp_keepalive_time yesterday but tonight's amdump failed again.

As the upper message doesn't come from mesg, but from the data channel (?): Would it help to increase dtimeout here?

Update: even with lower tcp_keepalive_time and increased dtimeout that DLE fails repeatedly.

What's interesting: Another DLE is twice as big and runs through fine.
The change of tapes seems to trigger that "Connection reset".

I am thankful for any hints as I should manage to come up with a full backup of those DLEs as an archive set until June, 1st ....

Thanks, Stefan

http://wiki.zmanda.com/index.php/Results_missing

In /etc/modprobe.conf:

options  ip_conntrack_amanda  master_timeout=3600"

I'll try that soon ...