[Mailman-Users] Downgrade/migration problem

[Lars Bungum]

On Wed, 2003-12-31 at 18:12, Stephen Gran wrote:
> On Tue, Dec 30, 2003 at 11:03:51PM +0100, Lars Bungum said:
> > On Tue, 2003-12-02 at 03:46, Stephen Gran wrote:
> > > Mail coming in to the lists and going out to users works just fine.
> > > Mail going to admins is accepted by exim, and passed to mailman, but
> > > mailman never delivers list-admin mail.  I see this in logs/error:
> > 
> > > Dec 01 21:38:06 2003 (15951) Delivery exception: bad marshal data
> > > Dec 01 21:38:06 2003 (15951) Traceback (most recent call last):
> > >   File "/usr/lib/mailman/Mailman/Handlers/HandlerAPI.py", line 82, in do_pipeline
> > >     func(mlist, msg, msgdata)
> > >   File "/usr/lib/mailman/Mailman/Handlers/Hold.py", line 152, in process
> > >     hold_for_approval(mlist, msg, msgdata, NonMemberPost)
> > >   File "/usr/lib/mailman/Mailman/Handlers/Hold.py", line 218, in hold_for_approval
> > >     mlist.HoldMessage(msg, reason, msgdata)
> > >   File "/usr/lib/mailman/Mailman/ListAdmin.py", line 145, in HoldMessage
> > >     self.__opendb()
> > >   File "/usr/lib/mailman/Mailman/ListAdmin.py", line 69, in __opendb
> > >     self.__db = marshal.load(fp)
> > > ValueError: bad marshal data
> > 
> > I'm seeing the same erros in my logs/error (log getting very large,
> > btw), and don't understand what's wrong.  Did you get any closer to
> > resolving this issue?  Everything seems to be working here as well.
> Some of the request.db's were corrupted, IIRC.  I had to do
> some for loop, echoing thefilename back out, to get it to tell me which
> one, then Imoved it aside and reran until it worked.  Let me see if I
> can find what I did . . . no, I can't find it right now.  Basically just
> loop over all of your request.db's while running the cron job's
> (chekc_dbs ?), and when it errors, you've found a corrupt one.  Move it
> aside, and rerun.  Rinse, lather, repeat.  It took a while, but I
> finally got all 50 or so running OK again.

(CC-ing to list as it probably is of general interest).

Thank you!  Your solution seems to have solved my problem.  However, I
just moved the corrupted request.dbs away and replaced them with empty
ones.  Is there any way to look into these databases to find out what
has gone wrong with them?

--lars