tinc detects itself on the VPN due to intermittent connection problems, and dies

[Xaquseg]

On 7/15/2011 4:19 PM, Guus Sliepen wrote:
> On Fri, Jul 15, 2011 at 04:06:59PM -0500, Xaquseg wrote:
> 
>> A relatively recent update (1-2 months) has introduced a fairly major
>> issue where connection issues seem to cause tinc to decide that there's
>> another node of the same name on the VPN, leading to it quitting.
>>
>> It seems that it's detecting itself somehow, through an older
>> connection, perhaps? The daemon stopping itself is a major problem,
>> though, as it has to be manually restarted.
> 
> That is indeed undesirable. Could you send me the log output from tinc when
> that happened, and also the log output after doing "tincd -n <netname> -kUSR2"?
> That would help me find out what is causing the incorrect behaviour.

I get a bunch of "tinc: Metadata socket read error for <various other
nodes>" during the connectivity issues, with some "tinc: Possible node
with same Name as us!" mixed in. After a bunch of those latter messages,
I get one followed by "tinc: Shutting down, check configuration of all
nodes for duplicate Names!", and tincd shuts down, as the message says.

I can possibly send you a USR2 output in a private email later, but I
will have to get permission to do so first.