On Thu 29 Mar 2018 at 17:20:58 (+0200), to...@tuxteam.de wrote: > On Thu, Mar 29, 2018 at 10:10:55AM -0500, David Wright wrote: > > On Thu 29 Mar 2018 at 10:46:17 (+0200), to...@tuxteam.de wrote: > > > On Thu, Mar 29, 2018 at 08:24:03AM +0000, Curt wrote: > > > > > > [...] > > > > Thanks for the ideas. I hadn't come across the virtual buttons. > > > > I think I'll look at what processes are running by the time it > > tries to mount /home, and which of them might involve looping on > > that timescale. > > If it's something in user space, it'd be interesting who has > the console open.
That comment made me remember about one possible candidate; I have @reboot /sbin/kbdrate -r 8 -d 500 -s in root's crontab as a matter of habit. But removing it had no effect. I can definitely only observe two CRs, one at the unlocking prompt (assuming I wait for it), and one at login. Of course, to be able to observe a third, I have either start typing my username or press <return> (in order to get the cursor out of column 1) and that very keystroke could be enough to stop whatever's happening. > > Then, when it happens again, I'll try a remote login to see if > > I can catch some process at it. It hasn't yet happened (on the > > three occasions) at an opportune time for me to investigate. > > > > I agree, the first character is probably a NUL. The second is > > probably CR. Both can be used to terminate strings in different > > situations. Or perhaps my termcap has fallen into a 1960s wormhole: > > anyone remember WRU, and the Here Is key, … > > Faintly. I also dimly remember about some serial interfaces > sending NULLs when idle, but I can't put my finger on that. My memory of these machines is of a roomful of them all bursting into life when they rebooted the PDP11 in town, two miles away. Then waiting for peace to be restored as their motors timed out. > The advantage of old age is that you tend to forget things :) Cheers, David.
