[Mailman-Developers] Update MM 2.1.2->2.1.5 fails and no help from mailman-users

[Frank Thommen]

Dear Mailman developers,

I dare to post this problem to mailman-developers, because it is
critical to us to solve it.  I assume that "code insiders" as you can
possibly help in this case.


While updating MM from 2.1.2 to 2.1.5 (Mailman stopped,  MTA stopped,
lock directory cleared manually, Apache stopped, 'check_perms -f' done)
the process of updating gets stuck at a specific point.  I added a
'print mlist' in 'def update_pending()' after
'for listname in Utils.list_names():
        mlist = MailList.MailList(listname)'

to know where it happens:

------
# bin/update
Upgrading from version 0x20102f0 to 0x20105f0
getting rid of old source files
Updating mailing list: XXX

  [...updates all lists with one "could not acquire lock" message...]

Updating Usenet watermarks
- nothing to update here
Updating Mailman 2.1.4 pending.pck database
WARNING: Ignoring duplicate pending ID: 7544.

  [..warns for many more "pending IDs"..]

<mailing list "list1" (locked) at 4b8300>
<mailing list "list2" (locked) at 4bbdf0>

------

...and here the process stalls forever.  I ^C-stopped the update several
times, but it gets always stuck at the same list.

If I nevertheless start 2.1.5, I get around 20 of the following errors;

------
# mailmanctl start
Starting Mailman's master qrunner.
# Traceback (most recent call last):
  File "/usr/pack/mailman-2.1.5-inf/bin/qrunner", line 270, in ?
    main()
  File "/usr/pack/mailman-2.1.5-inf/bin/qrunner", line 230, in main
    qrunner.run()
  File "/usr/pack/mailman-2.1.5-inf/Mailman/Queue/Runner.py", line 70,
in run
    filecnt = self._oneloop()
  File "/usr/pack/mailman-2.1.5-inf/Mailman/Queue/Runner.py", line 99,
in
_oneloop
    msg, msgdata = self._switchboard.dequeue(filebase)
  File "/usr/pack/mailman-2.1.5-inf/Mailman/Queue/Switchboard.py", line
147, in
dequeue
    data = cPickle.load(fp)
EOFError
[...]
#
------

After that, MM 2.1.5 does not deliver any mails.  I can go back to 2.1.2
and mail delivery continues (even for mails posted in the meantime, i.d.
when 2.1.5 was running), after having given me a about a dozen errors of
the following type:

------
Jun 26 21:43:27 2004 (1204) lost data files for filebase:
1088195523.0786779+4c5e6189f5cc4ba967f1556ffb911c03c84d6174
------

If I then go back to bin/update, the process stopps within the same
procedure as above, but at an other maillist (see above).


An other obervation is, that when doing bin/update, I get errors like
the following with some maillists:

------
# bin/update
[...]
Updating mailing list: XXXX
Updating the held requests database.
- updating old private mbox file
    looks like you have a really recent CVS installation...
    you're either one brave soul, or you already ran me
- updating old public mbox file
- This list looks like it might have <= b4 list templates around
Traceback (most recent call last):
  File "/usr/pack/mailman-2.1.5-inf/bin/update", line 782, in ?
    errors = main()
  File "/usr/pack/mailman-2.1.5-inf/bin/update", line 672, in main
    errors = errors + dolist(listname)
  File "/usr/pack/mailman-2.1.5-inf/bin/update", line 357, in dolist
    os.rename(o_tmpl, n_tmpl)
OSError: [Errno 2] No such file or directory
#
------

After that, bin/update exits and I have to restart it.  The next run it
will give the error with an other list.  This happens three times, so
the fourth time bin/update runs through and finally stopps at the error
described at the beginning of this posting.


I'm running Python 2.2.2 on a Solaris 9 host.


Do you need more information to track down this problem?

Please reply to my personal address (fthommen at inf.ethz.ch), since I am
not member of this list and please be very specific if it comes to
coding, as I dont really speak Python.

Thanks a lot in advance from me and my 70 maillists :-)

   frank


----------
Frank Thommen, IT Support Group, D-INFK, ETH Zuerich
Mail: fthommen at inf.ethz.ch; Phone: +41-1-63 27208 (Mon-Thu)
Web: http://www.isg.inf.ethz.ch
----------