Dale Miller

I really need to stop putting things off. I have been remiss in obtaining a tuxedo for the choir concert next Sunday. I knew there was a good place to order one online, but I just hadn’t gotten around to it. (Three weeks ago would have been about the right time.) I signed on this morning to submit my order, hoping I wasn’t cutting it too close. Lo and behold, their website ordering wasn’t working. So I waited around until the brick-and-mortar store opened and called them up. Apparently they aren’t taking mail/web orders this time of year because they are too busy. Okay, now what do I do? I know I could buy or rent one in town, but that’s just too much money. I found another web-store that seems nice, but they wanted $48 for shipping if I wanted 2-day. Grrrrr. I opted for standard shipping and will hope for the best.

I also decided to deal with the fact that I have a GMRS radio and had never applied for a license. I signed onto the FCC web site and realized that my amateur license had expired, so I submitted a renewal for that. At least that should be taken care of. I also discovered that a GMRS license is $80. That’s a bit too steep for me right now. I think I will just operate on FRS frequencies for the moment.

I actually feel rested today for the first time this week. I slept about 10 hours last night. That may have something to do with it.

The weather is cold this weekend. I was looking to sunlight and warmth. I’m apparently going to have to wait. It’s 55 degrees and overcast. :(

Alton managed to fine a backup of the test database that I inadvertently destroyed. It’s about a month old, but that’s substantially better than none at all. He finished that restore last night, and Peter was able to recover it today. That part actually went right. So, we were finally back on track to try the testing that we were supposed to begin Wednesday night. Peter decided to save time by copying the files off of Apollo rather than restoring them from tape. As he was about 3/4 of the way through, Lance deleted them all. When Guil presented the database clones to Gemini and I mounted them, I discovered that he had cloned the wrong file system, because it was mis-labeled in the SAN. Alton started a restore from tape, and the tape drive malfunctioned, holding onto the tape in the process. Just how many things can we have go wrong on what was supposed to be a simple test?

It’s Friday night. It’s been a long week. I’m going to go climb in bed and read a book.

I’m sleepy. That’s not unusual for a Friday morning this term, since I tend to stay up late on Thursdays due to choir. I’m just not hitting on all cylinders lately. I left my music folder at home yesterday, so didn’t have it for rehearsal. I pretty much have it memorized, so it wasn’t that big of a deal, but still a bit annoying. This morning I was most of the way to work before I realized I had forgotten my cell phone. That’s always inconvenient. It was especially so today, because poisonunicorn invited me to call her and I couldn’t. The university rather frowns on personal long-distance calls from the office phone.

I think Gemini is mostly back to a usable state. I still don’t know if we are going to be able to recover the test database. I’ll find out when Peter gets into the office in a few minutes. At least my part in all of that is at an end (I hope). Now I get to try to troubleshoot a routing problem between two of our exchange servers. Exchange is always a challenge because Microsoft does things their own way, and the diagnostic tools are on the minimal side. I have messages that tell me a delivery failed, but no explanation whatsoever as to why it failed. The SunOne stuff is only marginally better, though. I miss having a mail system that I could actually debug.

I whacked my elbow into a railing last night at rehearsal, and it is still hurting. It’s not really painful, but it’s enough to be a distraction.

I have this temptation to drive to Fort Worth this weekend. I would only be able to stay one night, though.

I started the O/S install last night before going home. I came in this morning and configured the network, loaded licenses, etc. I turned the system over to Alton who installed the Data Protector client and restored the backup copy of the root file system to a temporary directory. I then mounted up the old (now trashed) disk on /mnt and copied the files over. I shut the system down and changed the boot disk. Wonder of wonders, it started up perfectly. Okay, so things are going great now. It’s now time to actually do the restores that started this whole affair. We needed space to stick the data, and I had a drive that had junk on it, so I did the following

1) Log onto Gemini as root
2) mkdir /data/extra
3) mount extra#extra /data/extra (Yes, I know it’s strange syntax, but it’s advfs)
4) cd /data/extra
5) rm -rf *
6) realize that I had mistyped the directory name in step 4, and I was doing an rm-rf * against /. This not a good thing.

So, rebuild the system disk again, and hope that Alton has backups of the data directory that I also managed to destroy. This is not going as planned. But I know how to recover a TRU64 system from backups now.

We were planning to use one of our TRU64 systems to test disaster recovery procedures this afternoon. The plan was that we would restore a copy of the production system to spare disks on the test system, do a complete recovery of the production database, make sure we had all of the Oracle install, Banner install,etc. just as we would have to do if we lost the machine room and were installing from tape somewhere else. The assumption was made that on a new computer we had already installed the O/S, so we weren’t planning to test that part. However, as I was doing a SCSI bus scan to locate the new disks that Guil had presented from the SAN, the system crashed. That wasn’t nice. A SCSC scan shouldn’t cause a crash. Okay, so something glitched. No big deal, we’ll just reboot and keep going. Besides, that will be guaranteed to find the new disks. So, start the reboot and the system crashes with an odd error message. Hmmmm…. must be something about the new disks — have Guil remove them from the presentation — same thing. Turn the system off and back on, just for grins — no change. So, get HP on the phone to see what might be causing this problem. First-level tech had me try a couple of things with no result. Second-level guy was quite knowledgeable. He had me start up the system from the CD, mount in the old root file system, wipe out the device configuration, rebuild it while booted from the CD, and try again. Okay, got past that problem, but it appeared that the kernel wasn’t right for that architecture. What??? Could we be looking a the wrong system disk? I hope not, because that would mean we were messing with the production root file system, while it was running. A quick check showed that not to be the case — whew. Okay, so what was going on? Were we running off an old copy? No, we don’t have one of those — checked the WWID — yep, that’s the right disk, so how could the kernel have become screwed up. Just then Alton mentions “You know, when I started the restore, I think it wrote some of the files to the root partition rather than where I wanted them to go. I killed the restore and restarted it.” Yes, that was the problem. Alton had written Apollo’s system disk over the top of Gemini’s while it was running. Okay, we’re no closer to having a running system, but now we at least know what happened. So I started a fresh install of the O/S on a spare disk so we can restore the last backup of the good system disk. We’re making our disaster recovery testing far more realistic than I was intending.

Sleepy — headache — don’t want to go to work.

What else is new?

For the last several days I’ve had a craving for cake. So, when I was at Kroger tonight, I decided to look at the cakes at the deli. The ones I wanted were at least $12.00, and I thought that was just too much to pay for a cake. So I headed over to the cake mixes, and picked up a cake mix and pre-made frosting (yes, I know that’s cheating, but it’s been too many years since I have made either a cake or frosting from scratch). I realized that it had been so long since I have baked that I didn’t have any cake pans (lost in one of my several moves). I headed for that aisle and, of course, couldn’t make myself buy the $1.69 standard cake pans. Oh no, I had to have the heavy, Teflon-coated pans for $6.99 each. So, I came home and made myself a $20.00 cake since I didn’t want to buy a $12.00 one. But that’s okay, the next one will cost $4.00. I also came to the realization that I don’t have a cake carrier or lid, so I don’t have a good way to store it. It will be going to the office tomorrow, and hopefully none of it will come home. So then I will have spent $20.00 for one slice of cake. Yeah, that’s the way to save money.

I just finished watching “Sky Captain and the World of Tomorrow”. It was quite an enjoyable escape from reality. I had no problems with the anti-gravity, robotic war machines, a P-51 flying halfway around the world, etc. No, what bothered me was the camera that Polly was using. After inventing all the futuristic machinery, they used an Argus C-3 as the camera they would feature. It’s a lovely little camera, but they messed up, and it was a constant irritation to me throughout the film. The C-3 had in interesting quirk in that the film winding mechanism and shutter cocking mechanism were separate. In the movie, Polly never cocks the shutter. In fact in one of the early scenes you clearly see the shutter cocking lever in the released position as she takes a picture. There is also an indoor scene where she turns on the (non-existent) flash. The scenes through the viewfinder are wrong also, but that didn’t bother me quite as much. The real annoyance was the “film counter” that someone in Hollywood grafted onto the top of the camera. The C-# does, in fact, have a film counter, but it doesn’t look like that (it’s the little dial to the right of the big annoying one that’s grafted on).

However, since I had been noticing everything about the camera throughout the movie, the last line had me nearly dying of laughter. If you haven’t seen the movie, I won’t spoil it. It’s worth seeing just for that final laugh.

In case you wondered what to do with those leftover Victoria’s Secret bags …

They make good kitty sleeping bags.

The UALR mail server started acting up yesterday afternoon. It initially started by spitting errors to the console, and responding very slowly. In lieu of spending lots of troubleshooting time while the system was unusable, I decided to try the old “turn it off and back on” method (after a clean shutdown, of course). I did that, and there were no more errors, but things still ran slowly. I spent off and on most of today trying to troubleshoot it. I tracked it down to one filesystem, and I have started running a defrag on it (yes, it’s Unix, and yes, there is such a thing — it’s using VSFS). The percentage of time the processor is spending in iowait is down somewhat, but it’s still not where it should be. I really hate problems that don’t have an obvious cause. Just give me an error message, blink a red light, or go up in smoke. Then I know what to replace.

Sometime this afternoon, Jack had the idea to all go out after work to a place downtown. Apparently several people were going to go, and he urged me to. Jack has the best of intentions, and he does try to include me, but I’m just not comfortable in those situations. I think it’s knowing that everyone at the table besides me has someone to go home to. It’s okay at lunch, because we all go back to work. But in the evenings, I’m the only one driving home to a dark house. That ends up depressing me far more than it should. I really should try to participate in the things Jack plans, since the folks I work with are just about the only friends I have.

Yes, I’m feeling sorry for myself tonight. Can you tell?

I think I’m going to have to work some next weekend. Apparently the state auditors decided that we haven’t tested our recovery procedures sufficiently, so we are going to simulate a catastrophic system failure and restore all the production files to the test system and bring up production. Not really a big deal, but I really wish they would just announce to the campus that the test system will be down for a day or so and let us do it during normal working hours. But that would make far too much sense.

I want to take a drive to Fort Worth and visit poisonunicorn. We talked for a little while today. It’s too bad she doesn’t live closer. But at least Fort Worth is a reasonable drive (discounting the current price of gas).