[Prev][Next][Index][Thread]

balltown futures

i'm sending this to both the bmw and italian-cars lists
as it applies equally to both; Al Bower's original message appeared
on the italian-cars-digest.

Al Bowers <[email protected]> wrote:
>Doesn't this sort of imply that the group is ready for its own
>newsgroup?

i think that it is; i've been swamped by work (the first major
software delivery on a significant government contract) over the
past three months, so i've been unable to properly tackle this
matter.

>Or, richard, would a small contribution from each of us allow an
>upgrade to balltown to allow the non-digest version of the list to be
>resurrected (see why I wanted to avoid news.groups? :-).

gory details follow; non-computer geeks may want to skip this:

the list format was revised due to apparent breakage in the interaction
between our mail agent (a fairly current sendmail) and the
Domain Name Service (DNS): our mail queues were backlogged to the tune
of 400+ messages because sendmail couldn't get the information
it needed from DNS to send the messages over the net.  it is unclear
why the two programs weren't always talking to each other.  since
the bulk of the messages were from the bmw and italian-cars lists,
the finger was pointed (incorrectly in my opinion) at the two
lists.  since my opinion isn't the most important one here (i don't
pay any bills; i only produce lisp code and collect a pay check)
the solution that i arrived at was one that kept a usable form of
each list in being.  the likely alternative would have been suddenly
pulling the plug completely.  my solution has worked in that the
mail queue backlog has dropped from 400+ to between 10 and 20
messages in the past 24 hours; the lists are no longer "targets" of
the management.  i have deliberately set the digest-transmission
threshold to digests of 20,000 characters, which should cause two
to three digests a day to go out.  a digest is forced out each
evening about 10pm, in case the day's traffic has been light.
these parameters will no doubt be tuned over the next week or two.

the sluggishness of the sendmail/DNS interaction may well have been
due to an inadequate link to the outside world; plans are already
in progress to upgrade our link from a full time 19.2K connection
to full time 56K sometime in the next month or two.  whether this
will be good enough to allow restoration of the original format of
the lists is unclear at present.  any other action that might
deal with the sendmail/DNS issues is not to be looked for; our part
time sysadmin is very busy, and there are a number of projects
that are higher on his priority list than supporting the non
work-related mailing lists.

sigh,
  richard