Re: [Sks-devel] Re: Recon failure

[Yaron Minsky]

The DNS theory does so far seem like the best one.  I think a
filter mismatch would end up with a different error.  It would be
ideal to get the logs from the other side, since that might clear up
why the other side is terminating the connection.  Without that
information, it's a little hard to debug what's going on.

Yaron

On 9/24/05, Daniel Johnson <address@hidden> wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Jason Harris wrote:
> How often is your IP address changing?  Ask your other peers to
> add a cron entry like this:
>
>   # keep up with ...'s dynamic IP
>   0 * * * * touch /.../sks/membership > /dev/null
>
> for hourly updates.

My address is usually quite stable, I've had occasions where I held an
IP for two months.  The last change was on the 19th when I replaced
the UPS.

According to the wiki, SKS versions higher than 1.0.7 (at least)
reload their membership file automatically every few hours.  Since
most of the network is at least 1.0.9 (according to
sks_network_today.png), and the oldest is 1.0.6 (keyserver.hadiko.de)
I don't think that this is the issue.  That said, I've still got a
similar entry in my crontab from when I first installed SKS.  :)

I was chatting with a friend about this (okay, I was venting) and he
brought up a good question: Has the filter set been changed?  I
don't recall seeing anything about adding on to yminsky.dedup and
yminsky.merge.

Good news: Peered with ice.mudshark.org , and syncing appears to work.
Between mudshark and aaiedu I'm staying in sync.

Daniel Johnson
address@hidden
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)

iD8DBQFDNaqG6vGcUBY+ge8RArz2AKC/oBgbrAF/40dwYdoP76goJ9LBoQCgprXU
aIfbwgAcBjonX9EH4oktPl0=
=Vc/O
-----END PGP SIGNATURE-----


_______________________________________________
Sks-devel mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/sks-devel