« Or not | Home | Main screen turn on »


Mon 13 Sep 2010 by mskala Tags used:

I'm going to experiment with updating this entry in near-real-time as I work on putting together my new computer system. As of now (first posting) it's about 1:45pm. UPS, after claiming to have tried to deliver the package of parts at 10:23*, now says they will try again between 2 and 3. I plan to spend that hour waiting in the lobby to intercept the delivery person so there can be no excuse of "we tried to buzz you and you weren't there."

* I don't believe this claim because I was here at that time, they didn't buzz me, the buzzer is unlikely to be broken, and they didn't leave a "we missed you" card, as they would have if they'd actually gotten to the door and been unable to get farther. I suspect the driver took one look at the construction on Roncesvalles Avenue (which, granted, is scary), aborted the attempt, and told his** supervisors I hadn't answered. On the plus side, the customer service people I talked to subsequently were very polite and alleviated my anger a bit. The real test will be whether the delivery actually occurs, 15-75 minutes from now.

** Not a sexist assumption - I am talking about a specific individual known to be male; or at least, the customer service people told me they'd talked to "him" on the phone.

To set the scene: I am starting with a machine opal, which has a dead motherboard. I want to take the two 160G IDE drives from it and recover their data onto tetsu, the new machine I will build with the parts I hope to receive soon. It will have four 2G SATA drives. RAID is the order of the day: opal had RAID0 and RAID1 in different partitions, and tetsu will probably have some combination of RAID0, RAID1, RAID5, and RAID6. (Yes, there's a reason for mixing them, which I may go into as updates proceed.) I also plan to use LVM on tetsu to make managing the large amount of space easier. The current condition of the data on opal's drives is unknown, because at present I don't have any other machine that can read IDE drives. I'm hoping that that data is in good shape and it was only the motherboard that failed. Now (1:50) I'm going out to wait for the UPS delivery.

3:05: Waited in lobby until 3:00; no sign of the delivery. Went back to my apartment and was in the act of unlocking my door when I heard the buzzer going off, inside. Hurried back to the lobby, leaving the key in the lock because it was sticking and didn't come out easily, and I took delivery of the package at 3:02. I stepped out of the building to sign the chit; turned to go back in; reached into my pocket for my front door key; and realised that I had locked myself out. Fortunately, I was able to find another door that was unlocked.

Now, the first step is to get the new motherboard, CPU, and memory installed and see that tetsu boots with just those things, before I risk the hard drives.

5:00: Well, this is disappointing. On assembling the system and trying to boot, first of all I got nothing at all. Traced that to having plugged all the front-panel power switches and lights into the "third front-panel USB" header instead of the "front panel switches and lights" header. With that fixed, the power switch at least works to start the thing up, but it fails to POST and gives a beep code instead: groups of three mid-length beeps repeated indefinitely. I haven't found a key to ECS motherboard beep codes, but if it's standard AMI, that indicates a probable RAM incompatibility. I have $400 worth of brand new RAM here that is supposed to be the same standard that this board claims to support, and it would sure suck if that turned out not to be usable. I suppose I can try each module one at a time to rule out just having a bad module; also, it's possible the old RAM from opal may be recent enough to work in this board.

5:35: So far, nothing. I have not exhaustively tested every module in every slot, but it certainly appears that the motherboard will not accept any of the four DDR3 RAM modules I have. It only takes DDR3 memory in the first place, and all the other modules I have from other computers are at best DDR2 (which won't even fit in the memory slots). So it looks like there's little more I can do tonight.

My next thoughts are to (tomorrow) attempt to buy some DDR3 memory locally; and start taking steps to return this non-working memory to TigerDirect. On carefully examining the packaging, I note that it is not pristine: this is more wear and tear on the cardboard box than can be explained by my having opened it recently. So I suspect I was not the first to buy this package of RAM. Combined with the fun I had with UPS, this is not leaving me with a good feeling about TigerDirect.

Of course, this assumes that the RAM really is the problem. It'd be nice if I could find a more authoritative reference for the startup beep code than just Wikipedia and random Web BBS postings. One small bit of confirmation is that the beep code when I boot with no RAM at all, is the same as with any combination of these modules.

8:04: Took some time out for a Skype meeting with people in Waterloo; during and after that, I spent some time on the Net investigating the RAM situation. It appears that all DDR3 memory is not created equal. There are different clock speeds, which I knew about, but there is also the issue of whether the memory is "for Intel" or "for AMD." Some modules are for both, but many are not, and my modules are "for Intel" whereas my motherboard and CPU are AMD. This could arguably be called my fault, because the fact that it was "for Intel" was mentioned in the catalog from which I ordered. The other side of that argument is that I am a highly trained computing professional. If I could make this mistake and order the incompatible modules based on the information in TigerDirect's catalog, then so could at least 95% of their other customers (and it looks quite probable someone else did, since I appear to have been sold a package that was previously returned by someone else) and then it looks like maybe TigerDirect didn't disclose the limitations of their product as clearly as they ought to have.

Anyway, some people claim to have used these modules with AMD CPUs, but not with my motherboard and CPU in particular. So I'm going to chalk it up to an expensive lesson, if they don't accept a return at full price. (And not so very expensive anyway - I said $400 earlier but it's really more like $260 purchase price for the RAM, and I'll probably get a partial refund.) In the mean time, I'll go try to buy some "for AMD" RAM in Toronto tomorrow. And adventures from that point on will probably go in a new entry.

I'd really been hoping that by this point I'd be writing about RAID and LVM configuration...


It's okay, you're criticizing him. (Note there is no gender-free version of bogeyman, hangman, or Orangeman.) Axel - 2010-09-14 01:13
Ah, but your permission doesn't count, for obvious reasons... How about "executioner" for "hangman"? I'm not sure about the other two. Matt - 2010-09-14 08:30
"Executioner" covers a larger semantic field than "hangman", the latter being restricted to an increasingly obsolete technology. So the two are not quite equivalent. However, good work is still being done in Iran, I hear, and there was that splendid performance in Iraq four years ago. Axel - 2010-09-14 10:48
Well there wouldn't be a gender-free version of Orangeman because last I checked, membership in the Orange Order (like the Freemasons and the Knights of Columbus) is restricted to men. kiwano - 2010-09-14 19:49
The Freemasons have a bunch of spinoff and adjunct organizations that are co-ed or female-only. Presumably there are women with some sort of "Orange" affiliation who aren't literally members of the organization. I know I've heard a folk-song lyric that went "It was the biggest mix-up/The world has ever seen/Me father he was Orange/And me mother she was Green" and it would seem reasonable that that could be reversed. Of course that lyric also contains the solution, if one thinks there is a problem to be solved: construct the sentences to use adjectives instead of nouns. Note that that may be desirable anyway - e.g. the perception that calling someone "a Jew" is more likely to be insulting than calling them "Jewish."

As for "executioner," well, making it non-gender-specific also makes it cover a larger semantic field. That's kind of the point.

As for "bogeyman" I think the answer is probably "bogey." The OED has more citations, including earlier ones, for that than for "bogeyman" anyway; one could claim that "bogeyman" is an unworthy modern innovation independently of any gender issue. That's in its sense of "bugbear"; its sense of "one stroke over par, in golf" came a few decades later. Matt - 2010-09-15 10:02
I don't think that any of the spin-offs that would refer to their (female) members as Masons are particularly appreciated by regular Freemasonry. That said, Wikipedia mentions there being a parallel organization to the Orange Order for women, and its members are called Orangewomen, so I suspect that there's no gender-free signifier for members of an Orange Lodge, just signifiers for both genders (and the women's adunct Orange groups seem to be closer to equals with the men's groups than what I've seen for regular Freemasonry). kiwano - 2010-09-18 00:35

(optional field)
(optional field)
Answer "bonobo" here to fight spam. ここに「bonobo」を答えてください。SPAMを退治しましょう!
I reserve the right to delete or edit comments in any way and for any reason.