Thursday, December 03, 2015

Power Outage (vaguely planned, for once)

So repair guys are here to fix our power cable. In order to do that they need to take our power down. Estimated repair time is about 2-4 hours (there are "lots of obstructions" in the yard), which is probably more than the UPS can handle.

So, expect downtime. Probably within about 15 minutes of this posting to whenever (you can do the math).

Update 1pm: I'm a bit disturbed that they left without saying anything and there are no obvious holes in the yard, but the power is back and the Magic Portable Transformer Box is gone, so either it's all fixed or they've put on a much more clever (and much more hidden) temporary fix. I guess we'll see how this goes.

Sunday, November 22, 2015

1/2 Power Outage

Well, that's something I'd never seen before.

Lights and outlets in various rooms went out but others were just fine, and flipping the circuit breakers did nothing. Obviously not a power failure since we were still getting power. And luckily the outlets with all of the Computer Stuff were in the latter category. But, still, WTF?

Turns out, we had a 1/2 power failure. Evidently if you have coming in a neutral wire and 2 hot lines at 110V each of varying phases, and something munches on one of the hot lines, you still get Power but it's Weirdly Dysfunctional Power. Power Company people arrived and now we have our very own Personal Transformer to turn the one hot line back into something our house understands. And then sometime in the next week or two somebody will be back to do Actual Digging to figure out where and how the line got munched and what to do about it for the long term. Wheee.

(...note that Lambda/etc is on a UPS once again and so should cope just fine with the brief service interruptions that will inevitably occur while they're fixing this. Unless we get an interruption that's not so brief. Or whatever it is comes back and munches one of the other lines in the interim; that would be fun...)

Friday, October 02, 2015

trying out new bogon filter

So in the category of Things I Should Have Done Ages Ago, as a replacement for the bogon filtering I used to do (now useless due to the IPv4 address space being fully assigned), I'm now using the Spamhaus DROP and EDROP lists, which is their list of IP addresses that are ownz0red by Various Annoying People, which, in theory, my ISP should already be blocking, but evidently they're not -- counters show roughly 40 connection attempts in the past 2 days; most likely entirely to my mail server.

Meaning if, within the last few days you've suddenly found yourself unable to connect, that would be something to check (because that's a list you really don't want to be on).

Sunday, May 03, 2015

Debian 8 -- heads up.

So I just noticed that Debian 8 ('jessie') was released about a week ago (sort of fun to be exploring the package archives and, "hey waitaminnit, what is 'stretch' doing here?... holy crap...").

 Which means I'll most likely be attempting an upgrade real soon now. Perhaps even later tonight depending on what the release notes say about how nasty this one's likely to be.  Perhaps also accompanied by some yoyo-ing in the next day or two thereafter

Monday, February 09, 2015

IP address not resolving

There was a storm in Washington while wrog was in New Jersey, and the UPS did not bring everything back up.  This, combined with some other misconfiguration of the secondary DNS server, is causing lambda.moo.mud.org to not resolve.

Until such time as wrog returns to Washington to fix the problem, you can use the IP address, which is 209.181.94.35.

Sorry for the inconvenience.

Update (5:20pm — wrog): My nameserver is back, but until Rob fixes his, nobody will be able to find it. In the meantime, if you badly want a name rather than a number, try lambdamoo.fyigm.com.
Update (6:16pm — wrog): Rob's wasn't actually broken. Note to self (and also anyone else out there who uses openwrt and the overlaid squashfs+something filesystems that they use):
  1. Never edit anything in the /overlay directory; this does not do what you'd think. Not that I was actually intending to edit anything there. Which then leads to:
  2. Never keep emacs buffers open on the /overlay directory, or if you're prone to doing this (say because it's a convenient window on the List of Things That You've Changed from the Standard Distribution) then you might want to consider using uniquify.el. (of course this is probably all moot since I seem to be one of the 5 people left on the planet who still uses Emacs, but anyway...)
Anyway, it's all fixed now.

Saturday, December 13, 2014

Digging is Happening

After many weeks of nothing, with occasional bouts of marks appearing on the curb and then getting washed away  by subsequent rain, actual digging is happening.  A crew showed up yesterday and ripped up a part of the neighbor's driveway(*) as the first stage in getting an actual line to our house.  Apparently, some other  crew is going to show up early next week to finish the job.

(*) not a good week to be my neighbor:  Thursday's windstorm took out one of our really tall trees in the front yard, which fell exactly between the two houses.  Missed the temporary line but managed to total his boat...

Friday, October 31, 2014

Moved to new IP

We are now up at the new IP.  Still need to get the websites/etc moved and there are still annoyances on the new firewall that need to be addressed.  From your end, the miracle of DNS should take of everything and you should be able to just reconnect, but in case your DNS cache is as badly behaved as the one at xenon, feel free to try either lambdamoo.fyigm.com or 209.181.94.35 directly (ooo, looks like 3 people have connected in the time it took me to write this, so this is all apparently working...).

In other news, one of the Burial Service People stopped by yesterday.  They need to mark stuff up and may also need to get permits to tunnel under the neighbor's driveway, but we worked out a plan and that's all supposed to be happening for real next week some time.  We'll see.  Guy also randomly decided to replace one of the splices on the current line that was apparently filled with algae (things that happen with 40+year-old lines)

Update (7pm):  Ok, websites and pretty much everything else that matters are now back at various new IPs all in the 209 world.

Update (2am):  And now I understand why xenon's cache was behaving badly.  Evidently registrar glue A records (what you need if you have foo.com with nameservice coming out of ns.foo.com) can take a long time to expire -- I'm seeing lifetimes on the order of a week. Fortunately for lambda, being a (grand)subdomain under mud.org, which isn't moving, the glue is a a couple of levels down from where it usually is, and, thus, already taken care of.  But this is an issue for some of my other domains.