Over the past few years, our household has turned from a Microsoft Windows centric one, to one that is populated almost entirely by Apple Mac’s, along with iPhones, iPads and Apple TVs.
While over the years, the Macs have proved resilient reliable machines, our experience has not been without problems.
The only one that has been big enough to matter is the one that has affected my personal iMac. It’s a 27″ monster, with 16 gig of ram, and a 500 gig SSD (not factory standard!). This machine is super fast with its quad i7 processor and the SSD. Unfortunately, over the past three months or so it has become unstable. The symptoms started with random kernel panic’s (think Windows BSOD), and then moved forward to include drive corruption and the machine just failing to wake up.
Troubleshooting a problem like this can be a real bear! I started with checking the SSD for errors, and it found quite a few. After repairing those, things stabilized for a while, but then it resumed having problems. I then thought that it was perhaps an issue with Parallels (which I use to run Windows and various Linux distributions). I uninstalled and reinstalled Parallels many times, including doing the full uninstall process as documented here.
At this point, I began to run into corruption problems on the drive so bad that it was corrupted beyond a simple repair, and I had to restore the whole system (back up your systems kids!). Fortunately this was easy to do because of the Time Capsule that we use to back up our desktops and laptops.
I then decided, since most issues seemed to be drive related, to upgrade the firmware on the SSD. After downloading a bootable image and “burning” it onto a USB stick, I upgraded and the problem was solved!
For a few days. <sigh>
Continuing along the path of thinking that it was a Parallels problem, I downloaded VMWare Fusion. I tried to migrate the Parallels VMs over to it, but it failed again. This reinforced the idea that perhaps the problem was with Parallels.
I finally decided to “start over”. I wiped out the machine, reformatted the drive, and did a clean install of OS X Lion. Fortunately this was pretty straightforward and only took about 30 minutes, including the time to download it over the Internet. I started installing applications (leaving out Parallels) and everything remained stable. Hooray!
Until I installed the FitBit application. Within minutes, the machine Panic’d again! Okay, so clearly the FitBit application is the culprit! I uninstalled it, and four hours later, the machine went down again.
I decided to try to find a way to stress test the SSD to force an error to come up so I could get a replacement. After some searching around, I discovered the Mac’s have “built in” diagnostics. The newest ones have it in eprom (or something similar) and you can get to them simply by pressing and holding D (for Diagnostics – those clever Apple people!) at the chime. Unfortunately, it wasn’t on mine, so I had to figure out how to get into it from a Snow Leopard boot CD (these are what originally shipped with my iMac). After several false tries, I figured out that it was on disc 2. So to get the diagnostics to work, I just put in disc 2, rebooted, held down the D key and voilà, a few minutes later I was looking at diagnostics.
Now, let me tell you, these things look like they are from 1985… and I don’t care! They are obviously written to diagnose a computer, not to have a super looking experience. Here’s where things get interesting. I don’t have a wired mouse, and it turns out that the wireless one doesn’t work (I had a wired keyboard plugged in at the time, so I can’t say if a wireless would have worked or not – I’m guessing not), so I had to use keyboard shortcuts to navigate the app. It only took a moment to figure out tab and up and down arrow were my friends. I tried to check the “thorough” diagnostics flag, but couldn’t make it work, so I figured I would just run the normal ones. I kicked them off and sat on the couch for a few minutes and… BAM! An error came up! Oh happy day! I can prove that my SSD has a problem and get it replaced!
Wait, hang on a second, this error says “MEM” in it?!
So get this, after spending months thinking my SSD was unreliable, the problems were coming from one of the memory chips! Son of a…
To make an already way too long story somewhat short, I then had to uncable the iMac, pull out some memory chips, hook it back up (just assume this part happens a lot), run diagnostics again, find no errors, swap the memory chips (using the same slots to make sure I was testing the chips, and not another set of slots and chips), run diagnostics and it repeated the error – consistency! It is my friend! Finally, I pull one of the chips out, ran diagnostics, got a clean bill of health, put that chip aside, reinstalled the other ones, ran diagnostics, clean bill of health and…
Five days later it hasn’t had a glitch. OH HAPPY DAY!
Posted from my now healthy iMac.
2 thoughts