Over the last couple of days, various apps had been freezing. The first to go was, ironically, ChronoSync, which kept freezing when trying to back up my main working folder stored on an SSD inside a StarTech two-bay enclosure. Its error messages indicated that the source disk had become disconnected, but that disk still appeared to be working fine. Then other apps started freezing, until even Safari stopped responding. However, notifications of that external SSD being ejected improperly were infrequent, and didn’t coincide with apps freezing.
It was time to start up in Recovery and run Disk Utility’s First Aid on the offending disk, which not only remained fully connected during the checking of its 27 snapshots and the active file system, but returned no warnings or errors. Back in normal mode, though, it soon ejected itself improperly.
Any part of the chain from the SSD to the Mac could have been at fault. I’ve never liked the USB-C/Thunderbolt ports on this iMac Pro, as sometimes just adjusting the tilt of the display is sufficient to elicit disk ejection warnings. That has been one of its very few irritations over the three years that it has been my production system.
So I removed the whole chain from the port on my Mac, installed the wobbly SSD in a plain SATA case with a Thunderbolt 3 cable inserted into a different port. At first, that seemed good, then up popped the improper ejection notification once again. It looked like the SSD itself was at fault, so I replaced it with a spare 2 TB SSD which I had been keeping in reserve in the second bay of the StarTech enclosure.
With the old SSD carefully labelled as being suspect (at least), I formatted its replacement and gave it the same name before restoring from backups. In this case, as this external storage contains a mixture of my most active documents and many kept only for reference, it could have required two restore operations. In the small hours of every morning, Carbon Copy Cloner makes a full backup of the disk, which is supplemented by hourly Time Machine backups of my root working folder. Thankfully, I didn’t need to restore from the latter as no files had been changed since its last full backup.
Once my fresh SSD was fully loaded, it was time to turn on Time Machine automatic backups and daily CCC backups again. That only served as a reminder that you can’t fool Time Machine just by renaming a volume: its pane had trashed my previous lengthy exclusion list and added the new external disk to that list, giving itself nothing to back up at all. So I had to rebuild my exclusion list, which made the next backup a full one.
As it turns out, that works in my favour. Now I have a first full backup of my root working folder as restored from its CCC backup, together with all the older backups Time Machine made of that same folder going back for over a year. Even if my restored disk has lost a file or two somewhere in the process, that will still be available from those Time Machine backups of the original SSD. I can therefore let CCC make its routine daily backup of the restored disk, safe in the knowledge that its full contents remain in a separate backup.
Some lessons relearned:
- Even SSDs made by reputable manufacturers can fail long before they’re ten years old.
- Good complete recent backups are essential.
- It’s even better to have more than one backup of all important files.
- Local snapshots are a useful adjunct, but are no substitute for proper backups when a whole disk goes down.
- Disk problems often present with generic problems such as freezing apps and odd errors.
- Improper ejection of disks can turn out to have a more serious cause.