The following are some details of a system failure and recovery process that I’ve worked through over the last week. Even though things seem relatively stable now, I’m keen to get any thoughts on what may have happened and how a similar scenario could be avoided in future.
- 2019 MacBook Pro (Intel), with a lot of legacy applications and data from a previous (2015) MacBook Pro
- Updated from 12.4 to 12.5 - update process quit and machine restarted to recovery screen, with message stating that system would need to be recovered from a backup or reinstalled
- Time machine backup hadn't succeeded with a 'complete' backup since a year prior - fortunately all files were recently backed up - similar to the issue at Time machine full restore only shows "complete backups" - so the most recent restorable backup was about a year old (however individual files were newer)
- Ran Disk Utility first aid from recovery, no problems identified with disk
- Subsequently ran Disk Utility first aid on external Time Machine volume (a USB HDD), however stopped this at around the 48 hour mark, after the process became stuck on 'Checking multi-linked files'
- Chose not to rely on time machine backup, so also restarted machine in target disk mode and copied all documents, libraries and applications to a separate backup drive, over several hours
- Also attempted to clone the whole machine to another backup drive, using carbon copy cloner - but after about 24 hours of copying this failed
- Using macOS recovery, erased Mac, reinstalled macOS 12.5.
- Initiated migration assistant to migrate data from my time machine backup (using migration assistant, as opposed to restoring, I was able to transfer most recent files, not just those from the complete backup of one year ago)
- Reset PRAM, reset SMC
- Ran system diagnostics (boot+D), no issues detected
- Migration assistant failed (froze up) after about 24 hours
- Was able to boot to login screen, but noticed odd glitching - eg inversion of mouse pointer colours, cyan pointer in a black box
- Attempted migration assistant again, but process ground quickly to a halt
- Terminated the migration, re-erased machine, reinstalled macOS 12.5
- Manually copied all relevant files and applications from time machine backup, reinstalled applications, audio plugins, ssh keys and other required aspects of local environment
- Considered options for rollback to 12.4 as a critical piece of software was declaring itself incompatible with 12.5, but was ultimately able to install it regardless
- Stable system appeared to be up and running
- Rebooted and got an apparent freeze on reboot (no progress bar movement after about 20 mins)
- Attempted to reboot in safe mode but unable to do so
- Considered contacting Apple support, but thought it would be helpful to at least see at what point boot process was failing
- Rebooted with verbose output, and left running for a while
- Noticed that though the boot was taking ages, it was slowly progressing through lines reading
tx_flush: 1170: disk 1 xidetc
- Eventually (maybe after 30 mins) login succeeded, was back to desktop(!)
- Restarted again and this time restart was in normal time
- Did some more setup on machine, subsequent restarts have been fine
- Machine appears to be running smoothly
All up, this saga lasted about a week. Now wondering if I can trust my system, particularly as I seem to have had:
Issue 1: Failure of macOS update to 12.5
- Theory 1: maybe due to legacy drivers or applications on a very long established system volume
- Theory 2: maybe due to time machine volume plugged in during software update
Issue 2: Migration failure after macOS reinstall
- Theory: maybe due to time machine volume issues?
Issue 3: Very slow initial boot time after macOS reinstall and manual migration
- Theory: no strong theories - this did happen after a lot of software installation, so maybe some kind of related clean up
Another possibility, perhaps underlying each of these issues, could be some kind of underlying disk or other hardware error, but this seems unlikely due to clean bills of health from disk first aid and system diagnostics.
Needless to say, going through this process is something I’d prefer to avoid in future. For now, my main take-homes are:
- Don’t make any assumptions about the integrity of time machine (or remote) backups - check and verify these frequently
- Avoid having external peripherals connected during system updates (apparently - according to some online advice)
- Always ensure that time machine and/or online backup is complete, before even minor system updates
- Use a fast external SSD for time machine
Again, be great to get any thoughts or suggestions about managing these kinds of issues.