[btns] [Fwd: Update on the Mail List Issue]

[Julien Laganier <[e-mail]>]

FYI: Email messages sent to IETF lists during the 16 hours before Sept. 
5th 6:00 PM CEST were lost. If you had sent messages to the IETF mailing 
lists in that period, please check on the mailing lists' archives if 
they were indeed delivered, and retransmit in case they weren't.

Thanks.

--julien

--- Begin Message ---

I want to give everyone an update on what recently took place with the IETF
mail lists.

Several months ago Henrik wrote a new program to replace TMDA (which was
functioning until that time as the IETF spam protection mechanism). TMDA was
very inefficient and those inefficiencies led to server crashes whenever a
heavy spam run was made at the system.

At any rate, postconfirm has been functioning quite nicely until now --
thank you Henrik! -- but, as is the case with all new pieces of software, a
bug appeared.

The email delivery process generally works like this: email is sent to our
servers. The mail then goes to postconfirm so that it can be verified as
legitimate. If it is spam, the email dies there. Legitimate email is sent to
the mailman list servers -- Mailman is program the IETF currently uses for
email -- for processing.

Unfortunately, yesterday after a server reboot, the postconfirm program
failed to restart.  During the time that mail was lost, IETF mail transport
was up and it was passing email to postconfirm. But because postconfirm was
jammed, it reacted by throwing the mail away (rather than holding it and
retrying it later). Mailman, which was up and running, didn't receive any
list messages. And because of where the problem was, no one who sent a
message to the list would have receive a bounce message -- the message just
disappeared.

However, a little while ago Henrik located the source of the problem in
postonfirm and believes that he has repaired it. We also determined why
postconfirm didn't restart after the server reboot, and we've corrected for
that as well. Finally, we've assigned a watchdog program to postconfirm so
that if it ever crashes in the future, the server will start it again and
notify us that there is an issue.

While we believe everything to be resolved now, we appreciate your patience
as we continue to work to address any issues that may appear. And a big
thank you goes to Henrik, who just spent his Friday night working to repair
this problem. 

Alexa



_______________________________________________
IETF-Announce mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/ietf-announce

--- End Message ---

_______________________________________________
btns mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/btns