Lion Finally Installed (…and Here’s Why It Was Failing)

After a tremendous ordeal of trying to install OS X Lion on my January 2008 Mac Pro, I finally had a breakthrough. I discovered what appears to be a hardware incompatibility.

To properly tell this story we’ll have to go back in time.

In January of 2008 when this model of Mac Pro was available (version 3.1) it was definitely the cat’s meow. I bought a true boss of a system too. I picked up an 8-core 2.8ghz Mac Pro with 4gb of RAM. Later, I bought some third party RAM from Crucial to stuff it to the brim with 32gb.

I also visited Newegg.com to pick up three more hard drives. That was an easy decision. I picked up 3 more 750gb hard drives, all of them Seagate. When I was a PC guy Seagates or Maxtors were the only drives to buy.

I also picked up a Drobo. If you’ve read this blog for a while, you know what happened with the Drobo. To replace the Drobo I eventually purchased a Promise DS4600 and 4 Seagate ST31000340AS 1 terabyte drives.

The Promise unit had issues. It was finicky and liked to drop drives out of the array for no discernible reason. Oddly enough, the drive that failed out of the array most often was whatever drive was in bay 4. It didn’t have to be the same drive. You could literally swap the drives in and out of bay 4 and eventually it would fail. It was really bizarre. I opened numerous cases with Promise.

Promise finally came back with the information that this particular model of Seagate drives were not certified to work with the DS4600. Okie, I can handle that. No problem. I reviewed the compatibility list provided by Promise and selected a set of Hitachi 2TB drives.

I then moved on to fight the DS4600 again to make it work over eSATA. Only this weekend did that finally get resolved. But that’s another story for another day. If you want to hear it, let me know and I’ll be happy to tell it.

Anyway, back to the Seagate drives. After all of that drama and switching around, I now had a set of Seagate ST31000340AS 1tb drives… four of them to be exact. During this time I was also experiencing regular S.M.A.R.T. failures in the 750gb drives in the chassis of the Mac Pro. Those drives were being sent off and replaced on a fairly regular basis.

As these drives were replaced and rolled around, I decided that maybe I should move the 1tb drives into the Mac Pro and get 250gb of extra space. That’s kind of a no-brainer decision, right? I ended up with the original 750gb drive that shipped from Apple and three ST31000340AS drives in the other bays. I had the bright idea of creating a RAID-0 and installing Lion on it. I just knew it was going to scream.

…except that Lion wouldn’t install.

The install would always start off just fine. It would write files and then reboot. Then, somewhere in the next storage of the install it would just die. An error message popped up claiming that there was a problem and Lion couldn’t be installed.

how can this be? thought I. There’s no way Apple would release an operating system that has an incompatibility of this nature with a 2008 Mac Pro. this is insane.

I lost many countless hours of sleep to install attempts. I would try to install. I would watch it fail. I would research a little more. I’ve spent weeks trying to get through this. Nothing… and I mean NOTHING would get through with the install…

…until one time, it did.

Immediately, my trust level of the whole installation was suspect of course. Why would it fail to install so many times and then just out of the blue… it would work? It didn’t make sense. I had tried reseating hardware. I had tried pulling out the BlackMagic Intensity Card. I had tried pulling out the eSATA cards. I tried putting the stock RAM back in place. I tried everything. Nothing worked… until this time it did. Weird. It didn’t make sense.

I ran with Lion on the RAID-0 for a few days and happily thought I would go about the installation of Carbon Copy Cloner so I could set up clone tasks for the operating system disk.

cue music. It started to happen. Everything segfaulted. I could literally open the Console application and watch the crash reports roll in like a riot was going on in the Grid and reporters were on the air. No program was safe. Every one of them blew up. Sync your iPhone? Bam. iTunes died. Sync your iPod? SLAM. VTPCDecoder (or something) explodes. yeah, this OS is suspect.

I decided to whack the RAID-0 and try the install again on a single ST31000340AS drive. Guess what? The install failed… again and again.

I booked a Genius Bar appointment. Obviously, my logic board was bad.

I’m not sure what made me think to try it, but I did. One of the 1tb drives had died at some point and I received a replacement that was a completely different model. Also I ran across information on the net that a lot of people were having problems with ST31000340AS drives and certain versions of firmware. Those versions were SD1A and SD15. I looked over the drives I had. All of them had one of those two firmware revisions.

interesting.

I took one of the drives Seagate sent back that was a different model number. For the record, the drive was model ST31000528AS. I slapped it in the chassis and formatted it with HFS+. I fired up the Lion installer and hit go. I asked it to give me a full, fresh install of Lion on this disk. It worked the first time.

Not only did it work, it has been rock solid. Nothing is crashing like it was before on the other drives. Lion has become a joy to use the past two days. I stored away the Snow Leopard volume and kept it for emergencies.

I cancelled the Genius Bar appointment.

By now I think you can figure out what my conclusion. There’s something wrong with ST31000340AS Seagate drives. Don’t try to use them with Lion. Something about the kernel in Lion disagrees strongly with that model of drive. If you read around on the net you will find many, many horror stories with those drives.

Beware.