[Mailman-Users] "Lost" a list and now I'm stuck...
jonc at nc.rr.com
Tue Oct 28 09:56:13 CET 2003
On Mon, 2003-10-27 at 16:12, Simon White wrote:
> I had a system crash recently while a mailing was going out to a list
> that has about 11,000 members.
> In the mailman/lists directory the list still has its directory, but
> instead of the usual:
> config.pck config.pck.last request.db
> it has just these files:
> -rw-rw---- 1 mailman mailman 1670681 Sep 10 16:02 config.pck.safety
> -rw-rw-r-- 1 webusers mailman 29 Oct 22 01:44 request.db
> The list is no longer showing up in list_lists and check_db runs. It
> certainly existed before though!
> Can I still restore that list? I'm not sure I have all the subscriber
> details still... which would be a real shame to lose, it's a community
> site for new mothers (and fathers) for advice and stuff for babies,
> parents and young kids to do, etc...
> I had to clean out a load of rubbish in the qfiles to stop mailmanctl
> barfing a load of errors (Bad Marshal Data) ... sadly I don't have
> access to the full errors any more.
Have you run "check_db <listname> "?
Check a list's config database file for integrity.
All of the following files are checked:
It's okay if any of these are missing.
config.pck and config.pck.last are pickled versions of the
config database file for 2.1a3 and beyond.
config.db and config.db.last are used in all earlier versions,
and these are Python marshals.
config.safety is a pickle written by 2.1a3 and beyond when
the primary config.pck file could not be read.
Unless you have regular system backups of your list files, you should
also copy this file off to a safe location, before messing with it.
How much RAM does the server have? Are you using a lot of virtual memory
when sending out messages to this large list? You should look at your
systems resources and make sure they are sufficient to handle a list of
Take care - Jon Carnes
More information about the Mailman-Users