SMTP problem after recovering from disk full, clues?
A very old and well-used list has been fine for years on this box, running Mailman 2.1.10 on Red Hat 7.2. The disk got full and everything crashed. I was able to clear out a couple GB of space and get all services running again, but now mailman will not send any messages, either to the list admins or to members. Just before it died completely, admins got a "We're sorry, we hit a bug" email.
Incoming spam is properly discarded and logged in /logs/vette. Incoming posts show up in /logs/post showing 250 failures (there are 698 members on the list). The web interface seems to be functioning properly.
/logs/smtp-failure shows "Low level smtp error: (-2, 'Name or service not known'), msgid: ..." and then
May 19 22:02:15 2013 (16851) Low level smtp error: please run connect() first, msgid: ...
and then listings of each of the addresses where delivery was attempted, each with "delivery to xx@yy failed with code -1: (-2, 'Name or service not known')"
Nothing has changed with /etc/smrsh or sendmail, I can send and receive mail with mutt just fine.
When I reinstalled mailman and sent it a test post, I got the group mismatch error reminder in the returned mail message that I needed to run configure with --with-mail-gid=mail, which I did, now test messages to lists don't bounce back to the sender, they just don't get out of the server.
Following some of the FAQs I've looked with dumpdb and find no obviously corrupt .pck files. I moved ~mailman/lists and ~mailman/archives off somewhere else, reinstalled the same version, brought those directories back to ~mailman, and get the same behavior.
What am I missing?
-- David Josephson
I've noticed that if mailman is stopped/crashes/system reboot/etc. and there is anything in mailman/qfiles/maildir/cur that stuff seems to just hang in mailman until you stop mailman, clear this directory, and then restart mailman. May not be your problem but try checking it.
you might also want to run the various check/test programs in mailman/bin such as check_db and check_perms.
On 5/20/2013 1:45 AM, David Josephson wrote:
On 05/20/2013 05:20 AM, Richard Shetron wrote:
Maildir delivery to Mailman is experimental and rarely used. It is not related to the OP's problem anyway, as in his case, messages are being processed through the pipeline, and it is only the final SMTP delivery in OutgoingRunner that's failing.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
On 05/19/2013 10:45 PM, David Josephson wrote:
This issue can have several different underlying causes including things like permissions on /etc/hosts and /etc/resolv.conf. See the FAQs at <http://wiki.list.org/x/-IA9> and <http://wiki.list.org/x/AoE9> and some posts linked therefrom for various tests and debugging techniques you can use to help find the cause.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
On 05/20/2013 11:05 AM, David Josephson wrote:
Yes, but read the FAQs and do some of the tests outlined therein and in list posts linked therefrom to help you figure out what the problem is.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
On 05/20/2013 12:41 PM, David Josephson wrote:
Possibly you had previously set
SMTPHOST = 'localhost.localdomain'
or something similar in mm_cfg.py, and this got dropped reverting to the Defaults.py
SMTPHOST = 'localhost'
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
I've noticed that if mailman is stopped/crashes/system reboot/etc. and there is anything in mailman/qfiles/maildir/cur that stuff seems to just hang in mailman until you stop mailman, clear this directory, and then restart mailman. May not be your problem but try checking it.
you might also want to run the various check/test programs in mailman/bin such as check_db and check_perms.
On 5/20/2013 1:45 AM, David Josephson wrote:
On 05/20/2013 05:20 AM, Richard Shetron wrote:
Maildir delivery to Mailman is experimental and rarely used. It is not related to the OP's problem anyway, as in his case, messages are being processed through the pipeline, and it is only the final SMTP delivery in OutgoingRunner that's failing.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
On 05/19/2013 10:45 PM, David Josephson wrote:
This issue can have several different underlying causes including things like permissions on /etc/hosts and /etc/resolv.conf. See the FAQs at <http://wiki.list.org/x/-IA9> and <http://wiki.list.org/x/AoE9> and some posts linked therefrom for various tests and debugging techniques you can use to help find the cause.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
On 05/20/2013 11:05 AM, David Josephson wrote:
Yes, but read the FAQs and do some of the tests outlined therein and in list posts linked therefrom to help you figure out what the problem is.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
On 05/20/2013 12:41 PM, David Josephson wrote:
Possibly you had previously set
SMTPHOST = 'localhost.localdomain'
or something similar in mm_cfg.py, and this got dropped reverting to the Defaults.py
SMTPHOST = 'localhost'
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
participants (3)
-
David Josephson -
Mark Sapiro -
Richard Shetron