[Box Backup] Need to reload bbackupd conf to avoid timeouts?

Chris Wilson boxbackup@fluffy.co.uk
Thu, 2 Nov 2006 21:07:31 +0000 (GMT)


Hi Tobias, Ben and all,

On Thu, 2 Nov 2006, Tobias Balle-Petersen wrote:

> I have had this problem for a while. Here is an older log from the 
> client showing the error:
[...]
> Aug 24 11:20:44 yoiko bbackupd[11514]: TRACE: Deleted directory record for 
> /Volumes/raidA/files/guests/Bergen Bybane\n
> Aug 24 11:20:44 yoiko bbackupd[11514]: TRACE: Deleted directory record for 
> /Volumes/raidA/files/guests/Library1\n
> Aug 24 11:41:58 yoiko bbackupd[11514]: TRACE: Exception thrown: 
> ConnectionException(Conn_TLSReadFailed) at SocketStreamTLS.cpp(361)\n
> Aug 24 11:41:58 yoiko bbackupd[11514]: TRACE: Exception thrown: 
> ConnectionException(Conn_TLSWriteFailed) at SocketStreamTLS.cpp(426)\n
> Aug 24 11:41:59 yoiko bbackupd[11514]: Exception caught (Connection 
> TLSReadFailed 7/34), reset state and waiting to retry...

I have an idea what the problem might be. If bbackupd spends a long time 
scanning directories without sending anything to the server (which might 
well be the case if you have 1 Tb of data in mainly small files) then the 
connection could still time out. The KeepAliveTime relates only to sending 
keepalives while diffing a single file. It does nothing while bbackupd is 
scanning directories.

Tobias, I could come up with a patched version against trunk which does 
implement keepalives while scanning directories if you're willing to test 
it.

Cheers, Chris.
--
_ ___ __     _
  / __/ / ,__(_)_  | Chris Wilson <0000 at qwirx.com> - Cambs UK |
/ (_/ ,\/ _/ /_ \ | Security/C/C++/Java/Perl/SQL/HTML Developer |
\ _/_/_/_//_/___/ | We are GNU-free your mind-and your software |