|
|
|
|
|||||
|
||||||
From: Frank Reed (no email)
Date: Sat Apr 15 2006 - 21:01:06 EDT
George H, you wrote:
"What then caused that accumulation of old messages to be suddenly cleared
out, in a cloud of dust, and distributed months late, as seems to have
happened to many members recently? I have no idea. Perhaps some buffer-space has
been overfilled, and therefore flushed out, before starting to accumulate again."
Interesting speculation. I've been puzzling over this, too, in order to
define whether there is just one problem with the list, with several symptoms, or
several problems.
As you may recall, the same burst of echo posts followed by a list crash
lasting a few days occurred in mid-November. I went to the list archives to see
how that pattern matched the recent crash. Fortunately, one of my own posts
made it into the echo burst back then so I was able to do a little more
detective work. Here's the sequence of events for this one case: On October 18, 2005
around 4:30 in the afternoon, I sent two messages to the list: "Swinging the
Arc" and "Sextant Telescope Collimation". After a couple of hours, the first
had appeared in the list archive but the second had not. At least one other
list member DID receive the second message on collimation. I decided to post
a second copy of that post and, sure enough, a few hours later in the
archive, there is a copy of the message "Sextant Telescope Collimation". So what
became of the first post? I didn't worry about it. Now jump ahead to November
15, 2005. The list sent out a batch of echo posts, different ones to different
recipients (the ones I received in my inbox are not the same as the ones that
appear in the archive). And in the list archive on that date appears
"Sextant Telescope Collimation", apparently this was the original copy that had gone
missing a month earlier (and in the archive, it is marked with the correct
earlier date and time stamp). Right after that, the list crashed and was
offline for several days.
As you suggested above, this sounds like a bounced message box has been
dumped, suddenly flushing out messages that had been accumulated. But apparently
a side-effect of this process is to crash the list. Of course none of this is
supposed to happen.
The problem of bouncing e-mail on the Internet generally is a serious one.
It has become much more problematic within the past 12 months as more and more
systems have installed over-aggressive spam filters. This is certainly part
of the problem, and it is fair for Dan Hogan to "blame" at least that part of
the probem on the recipient. But there is still a problem on the listserv
side. Flushing the "bounced message box", for want of a better name, should not
crash the list --as it has now down twice in the past six months.
As for "NavList" (the google backup for this list which I initiated), it is
of course likely that it would be better at handling these sorts of problems,
but there is also a possibility that it would have some similar affliction
or worse (!). As I said from day one, the idea was to make that list
"experimental". It's time for the 50+ people who have signed up for that to start
experimenting. Any of the rest of you who would like to see how that goes, feel
free to sign up by sending a blank e-mail message to NavList-subscribe
[ATSIGN] googlegroups.com. This does not commit you to any permanent change. You
will receive a confirmation message asking you to visit a rather long URL. This
simply confirms that your e-mail address is active and that you intended to
sign up. After that you post to NavList [ATSIGN] googlegroups.com. And I
reiterate, if Dan Hogan joins, I would be happy to make him a list owner
immediately. We're talking about changing vessels, swapping a leaky boat for a
seaworthy one (still sitting on the ways), not changing captains. Aye.
As for solving the problem with the present list hosted on webkahuna, Ken
Muldrew stated the case clearly, so I'll quote him:
"The problems with Nav-L will certainly require Andy, and probably those
upstream of him, to do a lot of work if they are to be fixed. There is very
little anyone from this group can do to sort this out as it will require access
to the Apache configuration files and logs. A list owner doesn't have those
priveledges. Someone could ask Andy if he's willing to track down the bugs but
as an unpaid volunteer, but I doubt he will be too keen. Having done such
work in the past, I would need a lot of motivation; others, and perhaps Andy is
one such, would do it merely for the technical challenge. He is a senior
software developer and the problem may turn out to be easy for him to fix, but if
you have ever read The_Cuckoo's_Egg by Cliff Stoll, you can appreciate how
difficult it can be to track down apparently trivial problems with networked
systems."
-FER
42.0N 87.7W, or 41.4N 72.1W.
www.HistoricalAtlas.com/lunars
|