Press Enter for Maintenance

I have Ubuntu Mate 16.04 on a Raspberry Pi 3 B. For some reason, even previously good images are now stopping the boot sequence with
“Press Enter for maintenance
(or press Ctl-D to continue)”
Ctl-D seems to have no ill effects. What is going on?

Interestingly I’ve seen the same thing yesterday on a Linux Mint laptop.
Possibly a hardware failure? Have you tried a different SD card?

Yep, I keep a backlog of SD images, and previously good images copied to new cards give the same result

You should get clues as to what is wrong in the last system messages before the “Press Enter for maintenance” stuff.
If you don’t see the system messages on boot, hit F1 during boot.

Would it help anyone if I captured the booting stuff that precedes the subject message?

Yes it would, it’s hard to give too much information when reporting a bug. ^^

I’m on a Windoze system now, and can’t figure out how to send the whole thing, but here are the last few entries:
… Set up automount Arbitrary Executable File Formats File System Automount
… BSC1 Contoller at … (irq 83) (baudrate 100000)
… 12c /dev entries driver
… media: Linux media interface: v0.10
… Linux Video Capture interface: v2.00
… Mounted Debug File System.
… Mounted POSIX Message Queue File System.
… Started Journal Service.
Press Enter for maintenance
(or press Control-D ro continue):

Enter takes me to a hash prompt.
Ctl-D finishes the boot and starts the GUI normally.

… and there is a long pause - maybe 5 seconds - between the last log entry and the appearance of the Ctl-D thing.

When I hit ctl-D, there is a quick flash message:

sulogin cannot read dev/tty1 … something about not permitted

Sorry for the flood of info, but you did ask …

This time, the log entries are in a different order and the last one was
… random: nonblocking pool is initialized

The message that appears on the console after the “… to continue):” finishes with “Operation not permitted”

… and to complete the story, I swapped out the Pi processor, with the same result.

Well, not quite complete - I went back to an image generated on 5/24 and it boots flawlessly. There is still a 5 sec pause after the stuff quoted above, but this one then cranks up the GUI. The hardware configuration is identical to the one that failed with the SD image generated on 6/6. If I knew how to pull up the installed updates, we might be able to narrow this sucker down. I have been careful to install updates whenever they are available; looks like there’s one I can do without. Meanwhile, I am going to develop on this stable image until I hear something.