For the last few days, my Philips external monitor has not come on when I’ve booted up my MacBook Pro in the morning.
Oddly, I discovered more or less by accident that if I unplugged the HDMI to USB-C adapter and moved it to a different port, the display would spring to life. Ok … that’s at best a work-around, not a solution—as I indeed found out later.
First round of debugging
Things I tried in order to resolve the issue:
- unplugging and replugging the monitor
- unplugging and replugging the adapter
- replacing the HDMI cable
- rebooting the computer
- rebooting the computer in Safe Mode
- rebooting the computer to reset the SMC
- rebooting the computer to reset the NVRAM
- reinstalled macOS Catalina (a 55 minute operation on rural internet), and
- a few other things I can’t recall just now, but which I’ll add here if I think of them later.
Apple Support, take one
I contacted Apple Support under my Apple Care plan, and once I’d convinced the person at the other end that I’d tried all this, I was escalated to a senior adviser who had me do the following:
- delete two files of the form com.apple.windowsserver*.plist, and
- reboot.
He was confident this would solve the problem. He was wrong.
He had told me that if that didn’t work I could escalate to Apple Engineering, and he’d put a note on the case to that effect, but warned me I’d have to use the phone for that, rather than chat (likely due to the reboots I’d be likely to have to perform?). So I called in again.
Apple Support, take two
Asked to escalate to Apple Engineering. Whether that’s what they did or not, I don’t know. In any event, the new person reviewed the whole business with me and agreed that since this started happening right after the last patch to Catalina 10.15.7, it was likely an operating system issue and a restore to a previous point could prove that.
Unfortunately, the last time I contacted Apple (to get my input method for cuneiform running again), the solution, which involved resetting the SMC and NVRAM, inadvertently put my back-up system off-line, which I was unaware of until 10 days later when my system kindly informed me that I hadn’t done back-ups in 10 days! So this left with me only two options: backing up to October 25th (a long way back from ‘today’, which at the time was November 12th), or to November 10th; I opted for the latter as the less painful option.
Keeping in mind that I’d only just experienced a loss of back-up system, this was not the most comforting solution—and maybe a bit like dropping an atom bomb to kill a mosquito—but I didn’t like the idea of living with the problem, so in for a penny, in for a pound ….
Started the back-up. ETA to completion: 40 hours. Ok, so we know that’s likely an over-estimate based on insufficient experience, and likely to go down, but still.
As there was no point the support person sitting around for hours, she sent me an email so I could let her know when it was done and we could resume debugging, if and when. My options were to reply to the email with text or to click on a link in the email to go to an Apple messaging site where I could leave a voice message of up to 5 minutes. Good.
On my own
I monitored it regularly, with my heart in my mouth; indeed, the ETA dropped more and more rapidly, so it looked like it might finish around midnight. At 8 pm or so, I saw that it had recovered 605 GB or so and had 4 hours to go. Unfortunately, at 9, I saw the same thing.
This is when you have a serious decision to make, based on the undecidability problem: when an algorithm halts, you know it halts (and has either done what it should or not); but when it’s still running, it may halt in future, or it may not. We were in the second part of undecidability, and it certainly looked like the thing was stuck. So, decision time:
- cancel and restart, in the hope of a new restore not getting stuck, and getting done by morning, or
- leave it alone and hope.
I chose door #2.
Good call; it finished around 10 pm, at which point I set about the task of tidying up a variety of little things, including turning off automatic updates in System Preferences and automatic back-ups in Time Machine, as a precaution, so nothing would get clobbered inadvertently before I’d had a chance to confer with the Apple Support advisor.
Apple Support, take three
I dutifully opened the email from the Apple advisor and the first thing I noticed was that although she reiterated that I could just reply to the email, it actually came from noreply@apple.com. This was not encouraging, to say the least, but I sent an update anyway.
Then I clicked the link. I won’t go into the gory details, but over the next day and a half I have experienced
- a spinning ticker indicating that it’s stuck in la-la land
- a request to sign in with my Apple ID, following which I get a 403 Forbidden
- a successful landing at the message page, upon which I enter my updated information, press submit, and am informed that I cannot submit at this time, please try again later.
Repeat ad nauseam; the last is the least frequent occurrence.
Fortunately, she said she’d contact me next time she was in the office. Her office days being listed as Tuesday through Saturday, I expected a call Friday. I didn’t get one. I also didn’t get one today, Saturday (office hours start at 10:00 and it’s now 13:13).
Short version: there is no Apple Support, take three.
On my own, again
This left me wondering what to do next, as I was facing another undecidability problem, the algorithm this time being ‘response from Apple Support’. I decided to instal Big Sur, since going backwards had done no good ….
Step one: perform a back-up
Ordinarily, I take my life in my hands and just do it. But, having run out of ordinary feelings, I decided to do the Right Thing and perform a back-up.
Enter Time Machine, start backup. Back-up disc not found.
Say what? no, not the Hot-knife Boogie. The RAID drive was asleep. I haven’t yet figured out how to wake it up gently, so I did the violent thing and powered it off, then on again. And guess what ….
The external monitor came on.
Now, the external monitor is plugged into the same adapter as the RAID USB cable, so presumably there’s a connection (no pun intended) but I’m not sure what. It could still be some kind of software bug, or it could be the adapter, but that seems less likely under the circumstances.
Ok, so I do have a backup from yesterday and I’ve done next to nothing since then, so never mind the back-up, let’s go for the gusto and do the installation of Big Sur.
Step two: instal the new macOS
My system informed me the download of the new OS would take two hours. Instead, it took about one.
Installation took a half hour. Then some more time once it rebooted. So in the end, maybe two hours total.
And then nothing happened. It started the reboot, and I ended up with two blanks screens.
Undecidability #3: do I do a reboot with the power button or wait? Well, waiting has worked well so far, so off to lunch and we’ll see when I get back.
Nice lunch: toasted tomato sandwiches using my tasty kalamata olive bread. Yum.
Return, check computer: nope, still two blank screens. <Insert swear words here.>
Shitsky, as they say in Russia. Ok, guess we’ll have to do the power button thing. Well, I don’t know whether I touched the trackpad, whether on purpose or by accident, or whether grazing the power button had the effect, but in any case, both monitors came up, and I was dazzled by Big Sur. A better name for the theme might have been Big Disappointment, but I’ll let you follow the link if you want to know what I was thinking in detail.
What if I reboot?
So now the big question is: what will happen if I shut down now, and do a reboot from off? The RAID back-up is asleep at the moment, so it will be interesting to see what happens. I’ll be back in a moment to let you know. (Thank God for ‘Save draft’!)
This is what happens
This is totally weird. (Unlike, presumably, everything I’ve said so far.)
When I shut-down, the RAID back-up drive woke up. I’ve noticed this before, but had forgotten. I guess it just makes sure there’s nothing lying around it should attend to before we go down. Well, so much for rebooting with a sleeping RAID back-up; I guess we’ll see what happens in this scenario, and I can always do it again and wait for it to go back to sleep before rebooting.
The monitor on the laptop came on. The external one did not. This is like the recent days when I’ve had trouble getting the external monitor to display; in the past, if memory serves (could be a medium-sized ‘if’), both came on immediately. Bummer. Still have the problem.
Boot up slider bar progresses—seems to take quite some time—then the screen goes blank. What?? Oh, wait—it came back … and the external monitor came on. Okaaayyyy ….
I guess I’ll shut down and wait until the RAID drive goes to sleep, and try again.
This is what happens when I wait
Only the laptop monitor comes on. There’s no flicker and then two screens. End of.
Then I had another idea. I moved the RAID drive to that adapter from the back of an Airport Extreme, whose only function was to provide a port to the RAID back-up, after my interchange with Apple about the dead back-up system. The last person I talked to told me that my set-up wasn’t supported, and the one previous to that had suggested moving it directly to the Mac; so I did that.
It solved the back-up problem … but it could be that it introduced this one. So I decided to try something.
Testing another theory
I shut off the system and made sure the RAID drive was asleep. Then I unplugged the RAID drive from the adapter. Then I rebooted.
Presto! Everything is fine. When the system was running, I plugged the RAID cable back in. RAID came to life.
So this rather suggests that there’s some contention on the port and that maybe I need to put these things on different ports.
Fortunately, the ‘spare’ multi-adapter should be here within a few hours at most, then I can test that theory. Stay tuned!
Adapting to a new adapter
It seemed to me that there could be two potential issues leading to the port problem:
- Having the RAID drive and the external monitor on the same port is fundamentally a bad idea, and they should be split up, or
- The adapter had somehow become wonky and wasn’t passing through signals properly.
The second theory was the simplest, since it just involved replacing the adapter without moving things around, so I tried that first.
When I shut down, the RAID drive came to life, presumably to do some kind of clean up or other. Well, fine, let me reboot and see what happens. Reboot fine, both monitors came on! Well, may as well get some work done, do a back-up. So I did.
An hour or two later, I shut down again. At same later point I came back, and the RAID drive was still sleeping, so did a reboot. Monitor came on. Ok, looks good.
This morning, same thing. Monitor came on, RAID remained asleep. Now, does that mean that the RAID is now off-line? Easy to test: perform a back-up. Did that; no problem: RAID came to life and the back-up was performed, lickety-split, as nothing much had happened since the night before.
I also noted that my cuneiform .inputplugin, which had become dysfunctional (greyed out) over the course of yesterday’s activities, is now functioning just fine.
As far as I can see, case closed.
Post-script
Nothing is ever closed in the world of computing!
As soon as I actually started to do some work, I had to open MS Word because a student had submitted a .docx file rather than a PDF. Sigh.
It crashed upon load, repeatedly; the message suggested I try to re-instal. So I did. Another 1.2 GB song-and-dance, but it’s working again now.
For now.