A rough account of what I did with Emacs recently.
April 30
I can think of who to blame for this... wanting to manipulate an
Excel spreadsheet, I grab some python modules, and then spend a
half-hour fiddling around with them until I've got iterable
versions of some of the classes provided. Before getting back to
the task I'd initially set out to do.
April 28
While attempting to recover things from the dead harddrive I
found I was unable to unmount the disk. I thought about this for a
while and figured, boy, it must be really hosed. Then I discovered
my error: I was trying to unmount it from a shell that was using
the damned thing. Change directory, unmount, ta-da. I've tarballed
the files I might care about off the drive and will maybe see
about formatting it at the weekend on the off chance that this
fixes it. If not, toss & replace (the "toss" in this
case being " add to the pile of hard drives that I promise
myself I'll someday physically destroy").
April 23
Redid the WSGI stuff, and magically it worked this
time. Yesterday it was crashing, today it's fine. Tomorrow, no
doubt, it will catch fire and explode.
April 22
No messing this time: the box came up, and I unmounted the
busted drive until further notice. Had a look through the files
that were newer than the last backup; of course there's some stuff
I want to retrieve, but I think it can wait.
Performing various stunt feats with the RSS toy. Right now the
not-yet-pushed-to-my-webserver version is still Perl, using a
slightly hacked Sajax.pm for the AJAX bits, but I've rewired it so
that one of the AJAX calls actually goes via the django
replacement code instead. And much to my surprise it actually
works.
Putting the aforementioned hack on the webserver didn't go quite
so well; it would appear there are some version mismatches going
on. Grr.
April 21
AAAH. Server gets ~80% through a full backup - which looks like
it includes the busted disk - and then goes to black screen and
unresponsive mode. So perhaps unmounting the busted disk and
leaving it unmounted would be the smart thing to do
here.
April 20
Server back up and running, or at least limping. I did a restore
from the March 21st backup, since that's what was offered to me,
and as it suggested it might take a day to rebuild I left it be
overnight. This morning, the server was up on the restored image
without my having had to interfere, although it wasn't in wholly
happy condition. I've since established that the corrupted Server
HD has been logging intermittent errors since at least April 11th
without anything notifying me. Like, say, Apple's Server
Notifications. Said notifications also failed to notify me that
the server backups had not run since March 21st, either.
Also, it looks like server logs don't get backed up, or at least
not in a way I can go poking in the backups filesystem to find;
this is frustrating my attempts to find out when things actually
started going wrong.
On the plus side, the errant drive is actually mounted at the
moment, so I've been able to extract a list of files modified
since the restore point, and in theory I might even be able to go
simply copy off any that are important.
April 19
Ugh. The Mac Mini crashed; I hard-booted it, and within a
working day it had crashed again; after that it refused to
start. Booted into recovery this morning and found the System HD
was claiming to be irrepairably damaged. And of course the newest
backup I have is almost a month old, for what reason I don't know
- perhaps once again my system backups decided to silently stop
working, or started discretely excluding files I'd rather they
didn't... Of course, I'm fully expecting that it'll turn out that
some of the lost files are certs that I'd only just gotten back in
order after the last backup left me with a bunch of expired
ones. Le grande sigh.
April 15
Tried out Apple's prints-from-iPhoto service recently. Quite
impressed with the speed and the quality, given (for the former)
it's coming from Germany and (for the latter) I was working with
relatively low resolution JPEG images.
Backups: rebooted the server for other reasons, backups broke
again shortly afterwards - or never recovered in the first
place. Using the "disconnect all users, shut down time
machine, shut down file sharing" approach appears to have
worked this time. I should be able to script that, I think, once I
can figure out what the magic for "disconnect all users"
is.
April 14
Spent the weekend in Berlin, a birthday present from
Mrs. Waider. Watched The Wolverine, which was pretty
damned good despite some indications that it had been cut for
hotel viewing. Although I do want to know what the point of
showing Viper shedding her skin was, other than OH LOOK
CGI.
Also demolished Robert Ludlum's The Aquitane Progression,
which I somehow missed when ploughing through Dad's collection of
thrillers in my teens. Really quite the un-put-down-able
book.
April 9
Random backup failures again. Really, Apple, why do you have to
make this critical software so bloody
fragile? "Oh, the client didn't disconnect cleanly from the
server, so now you have to REBOOT THE SERVER before it can back up
again. P.S. Rebooting the server may completely block your backups
from working. Or not. We'll see."
April 7
Fiddling with some LiveJournal API stuff (handwave handwave);
the default polling interval suggested by the
"checkfriends" call is... ten hours.
The Guardian's RSS feed is pretty much useless at this point. A
shame, as it was a useful source of news that avoided BBC's whole
"scare quotes and nonsense questions"
approach.