Copius Free Time

Once again, my dining room floor looks like this:

My Dining Room Floor

My Dining Room Floor

Which can mean only one thing:  Someone’s having computer problems.

I’m generally pretty open to helping just about anyone with their computer problems and this particular client only comes to me when something is VERY wrong and not often at that.  In this case, her PC simply failed to boot.  It told her that it couldn’t find an OS to load.

The setup:  she has two identical SATA drives in RAID-1 using the onboard SATA-RAID controller.  The controller BIOS detected only one drive but refused to boot from a “degraded” array.  No matter how I plugged in the assumed good drive, BIOS refused to detect it as a SATA device.

So I took the machine home, got out a brand new SATA cable, and attached it to the only SATA-capable PC in the house that (a) didn’t cost $2600 and (b) wasn’t 100% mission critical.

The drive dangling out of the machine

The drive dangling out of the machine

The silver case with the side panel propped up in the left of the shot is the machine I’m using to test the drives.  The drive leaning against it with the SATA signal cable attached is the assumed good drive.  The drive cage and black case in the right of the shot is the machine that has the problems.  The first is that a retarded monkey superglued the SATA connectors into their sockets on the motherboard.  Oh, wait, I was that retarded monkey.  Four years ago, when I built this machine for her, I got tired to the SATA cables falling out of the mobo connectors, so I superglued them.

The Server, a brand new SATA drive, an assumed dead SATA drive, and the test PC

The Server, a brand new SATA drive, an assumed dead SATA drive, and the test PC

The black not-currently-rackmounted rackmount case is my server, dustpuppy.  The clamshell packaging on top of it is a brand new 160GB SATA drive for the server in case I lose a disk.  It might also come in handy to test the SATA controller on the malfunctioning PC’s motherboard.  The drive to the immediate right of the server is the assumed dead hard drive that came out of the bad PC.  When I connected that drive to the test PC, it wouldn’t even POST – the BIOS init locked up at the IDE probing.

Salvage Operation Underway

Salvage Operation Underway

The first order of business was to salvage the data on the assumed good drive before anything else happens to it.  Since I don’t have 500GB of offline dedicated storage laying around, I’m using the 3TB RAID array on the server to store a backup image of the filesystem on the drive.  I’m doing the tried-and-true brute force backup method: cat /dev/sdc1 | gzip > /net/dustpuppy/export/home/homelan/tmp/auntsueNTFS.gz

The image slowly grows

The image slowly grows

When all is said and done, I’ll be able to report the following:

  1. One drive is dead.
  2. Both SATA cables are dead (I had to cut the connectors out of the motherboard sockets).

And also whether or not the SATA controller (or the sockets) on the motherboard is dead, too.  The good news is that one disk is good so to get back up and running in the worst case would take a $25 PCI SATA HBA.

Advertisements
  1. #1 by baleon on December 14, 2009 - 11:21 AM

    So Josh, I’ve been meaning to ask you how you go about organizing all your computer parts. Mainly because my father is a software engineer and has a tendency to have multiple computers, and spare parts up his ass. He takes computers apart and rebuilds new ones. He takes over 3/4 of our basement storage area, and his computer room is on the other side of the rec room (which we can’t use because his computer area over flows even on the good days.) If I were to go in there and organize it behind his back.. where would I start?

    • #2 by Joshua on December 14, 2009 - 11:31 AM

      Lol my mom wants to know the same thing. Sorry to say I think it’s hopeless. I’ve worked in four different IT shops and all of them have work areas that look exactly like that: full of spare parts and half-assembled machines.

      All my parts are crammed in a drawer under my audio equipment rack. It’s hard to find anything I need when I need it. I’d much prefer a heap of stuff that I can just look at and go “Oh, there it is!”

      There are two ways to organize ADPE components: By Function and By Form. The By Function is more intuitive for him but wastes space and no two techs can agree what “functional group” a given part should go in. Example: I have a couple SCSI HBAs, some SCSI hard drives, some SCSI cables, etc. And I have some IDE hard drives and IDE cables. And then I have some SATA hard drives and SATA cables. Do I put them all together and call it “data storage?” Then why don’t I put my DVD+RW drives in there also? Or do I stack up the three different types of hard drives, the three types of cables, the three types of adaptors, etc. and label one “hard drives,” one “cables,” and one “HBAs?” Then when I need to replace a SCSI drive, I’d have to sort through the hard drives looking for a SCSI connector, whereas if I had all the SCSI stuff together, then all I’d have to do is look for any hard drive in the SCSI pile. You see what I’m saying?

      There’s no “optimal” way to organize computer equipment. Any form of organization will be a detriment to your dad’s ability to find the parts he needs at any given time.

      But don’t feel bad – geeks have argued this with their families, bosses, spouses, etc. since modern electronics came of age in the ’50s. The worst thing to say is “why don’t you get rid of some of this stuff?” Because invariably it’s all useful in some way and we might need it someday.

      • #3 by ale8oneboy on December 14, 2009 - 4:33 PM

        I couldn’t agree more. I find myself in the same situation. If you get rid of something, you or someone else is going to need it sooner or later. You could get rid of 30 pin memory. Then someone would craw from under a rock and say “Hey I could have use that for a memory upgrade!” Category (function), Sub-Category (type) is how we sort things at work. – mattkendrick.com

        • #4 by Joshua on August 17, 2010 - 7:40 PM

          Lol – agree.

  2. #5 by Joshua on August 17, 2010 - 7:42 PM

    I’d like to point out an unncessary use of cat in the above. The pipeline cat /dev/sdc1 | gzip > /net/dustpuppy/export/home/homelan/tmp/auntsueNTFS.gz could be rewritten as gzip < dev/sdc1 > /net/dustpuppy/export/home/homelan/tmp/auntsueNTFS.gz. It’d be faster and wouldn’t use a process slot for the cat instance. Not that a few cycles make a huge difference.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: