The time my computer almost died
I have a 2011 MacBook Pro. It’s old, but it’s also good. Until very recently, no newer models could take more RAM (I have 16GB, mostly because the incremental cost over getting 8GB was so low). Newer models do have higher res screens, but I really don’t care. One place where newer models completely fail is in the storage department. The new 15″ MacBook Pros may finally be interesting to me but I currently have a 13″ system and those ones did not get the same upgrades.
My computer came with a 128GB SSD and a CD drive. Within days of getting the machine though, the optical drive was removed and an Optibay put in (with the optical drive relegated to an external enclosure). I put in an extra 500GB of storage for a tiny fraction of what SSDs cost at the time.
Over time, the amount of things I could fit on the 128GB boot volume seems to have gone down. There has been a gradual movement of things to the hard drive (besides the obvious long-term non-speed-dependent things that started there). The computer is still mostly snappy, but some things are slower just because they rely on files on the hard disk.
Anyway, everything has been great with my setup for a very long time. That all changed a month ago.
I went to turn my computer on and since I’d put it to sleep at <5% battery, it had to resume from hibernation. Except, that it couldn’t resume from hibernation. Nor could it boot normally. I used a boot CD to try and repair the disk, but it reported unrecoverable failures. Not good.
The problem seemed to be the SSD, not the hard disk, so I connected it up to my Linux server to poke at it. It didn’t seem as dead as it had in my Mac, but it did seem to have problems. The sheer age of it, and the fact that it’s spent most of it’s life close to full made me suspect it had succumbed to the inevitable death-by-inability-to-erase that comes for all SSDs eventually.
Unfortunately, while trying to do a disk check (fsck.hfsplus /dev/sdb1), I accidentally ran the create volume command instead (mkfs.hfsplus /dev/sdb1), which wiped the SSD clean. So much for recovery then… good thing I have a backup!
My backup solution is a USB 2, 750GB WD disk I got from my sister. She gave it to me after it broke, but I figured out the error was a small one (damaged bootsector) and I’ve used it as my backup disk since (she didn’t want it back). My last backup was a month earlier but there’s not too much that changes in a month on my system and of the things that do change, most of those have external copies that can be used to being the system up to date.
Since I thought it was the SSD that was at fault, I grabbed an old hard disk I had lying around and shoved it in my computer. Restored from backup and… I was up and running again. Slowly, but running.
I had been wondering for a while now if it might be worth replacing the 2 disks with a large SSD. Compared to 2011 prices, SSDs are significantly cheaper now and it would simplify things, while allowing me to speed up the things that were slow due to being on the hard disk. It would also free up the OptiBay to put another disk into my Linux server (which is actually an old PC laptop).
I went ahead and ordered a 1TB SSD (Samsung 860 EVO). It arrived, I put it in my machine and… nothing. In fact, it seemed to have the same symptoms that the original SSD was having. Bugger.
Something I had read online while investigating my problem was that the HD (ribbon) cable in the MacBook Pro is susceptible to failure. It is a little loose and able to rub against the case, which is machined aluminium that isn’t very smooth. However, I found only a single report where someone’s cable worked with some disks and not others. Maybe this was my problem?
Apparently while investigating my original problem I had never actually tried running the SSD in the Optibay (I’m sure I had tried just about every other combination). But I went ahead and tried that with the new SSD and it worked. So I moved all the data from the 2 disks onto the 1 SSD. I also ordered a new HD cable, since I would still like to free up the Optibay.
One of the annoying things about Time Machine (Mac Backup) is that moving a file means it has to be copied to the backup disk, even though it’s already there. As part of moving all my files to 1 disks, I did a lot of “moving”, so I was expecting my first post-consolidation backup to take a while. What I was not expecting was for my backup disk to die!
It turns out, the backup disk must have been on its last legs. The stress of erasing all the old backups to make room to re-copy several hundred gigabytes was apparently too much for it and it went into permanent failure mode (ie. completely unable to work). All I can say is I am so glad it failed now instead of a few weeks ago!
Today I got a new backup disk. Apparently the 2.5″ bus-powered state of the art hasn’t moved forward nearly as much as the 3.5″ world has because there were no 14TB monsters at the shop. There were 1, 2 or 3 TB disks, with Seagate 2TB disks on sale for the same price as their 1TB disks. So I have a new 2TB backup disk. It’s got USB 3 but… apparently my computer is too old for that (it only has USB 2 controllers).
I can apparently get a Thunderbolt to USB 3 adapter (Kanex Thunderbolt to eSATA plus USB 3.0 Adapter) but it’ll cost as much as the backup disk… not sure I care enough to bother. I normally just let it run overnight to do backups.
So at the end of the day the tally is:
- 1x 128GB SSD maybe dead (if not, I might put it into my gaming PC, which has a 32GB SSD right now)
- 1x 500GB hard disk (will join my pile of old hard disks, unless I can put it into my Linux server)
- 1x 750GB USB 2 hard disk (dead)
- 1x 1TB SSD (in the computer)
- 1x 2TB USB 3 hard disk (still doing the initial backup)
- 1x HD ribbon cable (currently unused)
- 1x new HD ribbon cable (waiting to be tested)