On Wed, Mar 27, 2013 at 06:50:45AM +0000, Chris Maciejewski wrote:
> I am running tinc (1.0.19) on a cluster of around 40 servers (Xen
> virtual machines in two geographical locations). Overall it works very
> well, however yesterday it crashed on two machines in odd way. I was
> still able to ping VPN IPs of other hosts from the two affected
> machines but "vpn" interface was not visible when I run ifconfig.
> There was also no "tincd" process running (at least ps aux
wasn't
> showing it). They logged the following for approx. 20 minutes in the
> /var/log/syslog:
>
> Mar 27 01:52:14 us16 tinc.vpn[9528]: Flushing meta data to
<peer_host>
> (xxx.xxx.xxx.xxx port 655) failed: Connection reset by peer
>
> and eventually:
>
> Mar 27 01:53:03 us16 tinc.vpn[9528]: send_meta() called with NULL pointer!
>
> Any suggestions what could be the reason for this crash and if
> upgrading to the latest 1.0.20 might prevent it happening in the
> future very much appreciated.
I don't know the reason for this crash. Could you send me all tinc's log
messages from that day to me privately? That would help me debug the problem.
I do not think upgrading to 1.0.20 might help in this case. You might want to
try 1.1pre6 though, since that is different enough that this bug might not
appear.
--
Met vriendelijke groet / with kind regards,
Guus Sliepen <guus at tinc-vpn.org>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL:
<http://www.tinc-vpn.org/pipermail/tinc/attachments/20130328/59999ad3/attachment.pgp>