U4 wouldn’t start up from standby with remote.

[sub3R]

This one is unusual & unfortunately I didn’t have logging enabled. It happened yesterday afternoon (15-Mar-20).

Sequence of events prior to & after not being able to start the U4 from standby with the remote:

• My better half was watching a recording on our U4_02’s internal HDD (the other Ethernet LAN connected U4_01 was in deep standby & had been since the early hours of the morning).
• The recording being watched on U4_02 was deleted; it was switched to live TV & then switched to standby. (I have strict instructions not to switch to standby or deep standby when the File Player List is displayed because I think this has been the cause of U4_01 not previously shutting down to deep standby).
• Some hours later (a guess) I was asked what was wrong with the same U4 (U4_02), because it wouldn’t start up from standby.
• I tried the same remote & confirmed that it wouldn’t start up. Thinking it could be the remote batteries I tried the U4_01’s remote & got the same result.
• I then noticed the front display on the U4 showed the correct date but the time was slow by over an hour (it showed 16:18 & remained frozen at this).
• Trying to access the U4_02 via OWIF wouldn’t work but I found I could access U4_02’s File Player List from U4_01 over the LAN using the yellow button in the File Player List & play a recording.
• The last timer (autotimer) recording made on the U4_02 was ‘Fishing Australia’ on BOLD which started at 3:29pm & finished at 4:18pm that afternoon (including pre & post padding). The after event for this timer was ‘auto’ (‘standard’ in autotimer).
• I switched the U4_02 off & on at the rear switch to get it working again.

Some information I later discovered:

• The progress bar for the above ‘Fishing Australia’ doesn’t have the start & end markers for its time slot (the progress bar only shows -0.00 at the left & +48.36 at the right, which is the duration of the recording). When playing back, the show started at -3.11 from the beginning of the recording (as indicated on the left of the progress bar) & finished at +14:41 from the end of the recording (as indicated on the right of the progress bar). The recording is clean at the start & end without any obvious breakup. Skipping works fine.
• The next show due to timer (autotimer) record ‘Sydney Weekender’ on Prime7 HD from 5:25pm to 6:14pm didn’t record due to U4_02 being in this non responsive condition.
• The timer information for ‘Fishing Australia’ & ‘Sydney Weekender’ doesn’t look right to me (see OWIF screen captures)

Both U4’s are running 19.3 20191106 f/w with OverlayHD 1.70 plug-in & AutoMute plug-in. Both have a ‘Bluey’ USB tuner connected. Both are Ethernet LAN connected. Both use only FTA EPG. U4-02 connects directly to the TV via HDMI.

I have attached all the logs & files that were available along with some OWIF screen captures.

timers.xml.txt

(41.59 KiB) Downloaded 37 times

messages.txt

(77.75 KiB) Downloaded 34 times

enigma2_execution.log.txt

(13.75 KiB) Downloaded 40 times

dmesg.txt

(23.42 KiB) Downloaded 36 times

autotimer.xml.txt

(4.78 KiB) Downloaded 40 times

20200315 1529 - BOLD Port Macquarie - Fishing Australia.ts.sc.txt

(1.26 MiB) Downloaded 35 times

20200315 1529 - BOLD Port Macquarie - Fishing Australia.ts.meta.txt

(359 Bytes) Downloaded 36 times

20200315 1529 - BOLD Port Macquarie - Fishing Australia.ts.cuts.txt

(12 Bytes) Downloaded 44 times

20200315 1529 - BOLD Port Macquarie - Fishing Australia.eit.txt

(264 Bytes) Downloaded 35 times

[MrQuade]

If the network works, but the GUI and Web Interface have stopped, then it means the enigma2 process has crashed and not recovered.

You can simply cycle power to the U4, or restart it from a SSH session if this happens.

As for the cause of your crash....dunno. Maybe messages will show something. If not, make sure logging is enabled and wait to see if it happens again.

[prl]

Of the logging information posted:

The messages.txt and dmesg.txt files are ephemeral and each full reboot of the system clears them, so if the system (rather than the UI crashing with the black-and-white traceback information) crashed or was powercycled, they'll only have information from after the restart, not from the time of the crash.

The .ts.sc file is a binary file of structure information about the recording. It's moderately large, and not really all that helpful for debugging without the (usually very large) .ts file and the rest of the metadata files.

The fact that the clock had frozen for an hour indicates that it's likely that the UI had gone into an infinite loop or did something else that stopped the normal event dispatch loop from running. That loop needs to run for the box be able to react to the remote or update the front panel clock.

The debug log might help with that (if there are any debug prints in the infinite loop), but it may simply show the UI running normally and then just stop working.

[Grumpy_Geoff]

It may have been a driver crash, taking out enigma2 as it didn't exit -

Code: Select all

Sun Mar 15 07:23:11 AEDT 2020: executing enigma2
Sun Mar 15 08:14:14 AEDT 2020: enigma2 terminated with code 1
Sun Mar 15 08:24:42 AEDT 2020: executing enigma2

[prl]

It may have been a driver crash, taking out enigma2 as it didn't exit -

Code: Select all

Sun Mar 15 07:23:11 AEDT 2020: executing enigma2
Sun Mar 15 08:14:14 AEDT 2020: enigma2 terminated with code 1
Sun Mar 15 08:24:42 AEDT 2020: executing enigma2

Yes, if some call to the driver stalled in the main dispatch loop. A lot of the "action" in recording and playback happens in threads other than the main dispatch loop thread. If one of them hung, it would affect the recording or playback, but shouldn't affect UI interaction.

It's hard to tell without a debug log, an even a debug log may not be all that illuminating.

Something in the main dispatch loop seems to have gone out to lunch, though.

[sub3R]

... or restart it from a SSH session if this happens.

So will ‘init 3’ do that? I did think of it but wasn’t sure.

I’ll have to re-enable logging.

[sub3R]

The messages.txt and dmesg.txt files are ephemeral and each full reboot of the system clears them, so if the system (rather than the UI crashing with the black-and-white traceback information) crashed or was powercycled, they'll only have information from after the restart, not from the time of the crash.

I suspected that messages and dmesg may be useless.

One bit of information I forgot to mention; when I pressed the ‘Power Off’ button on the Harmony 650 remote, the red LED power indicator on the front of the U4 flashed each time I pressed it, but that’s all it did. It didn’t respond to the ‘Power On’ or ‘Power Toggle’ buttons.

Thanks for the explanation.

Something in the main dispatch loop seems to have gone out to lunch, though.

And seemed to have happened when the ‘Fishing Australia’ recording finished recording.

[Grumpy_Geoff]

Something in the main dispatch loop seems to have gone out to lunch, though.

Yes, that's why the ‘Fishing Australia’ timer is bereft of start/stop entries, as whatever may have been present for it in the in-memory table "disappeared" when the "gone to lunch" sign was put on the door.

[sub3R]

It may have been a driver crash, taking out enigma2 as it didn't exit -

Code: Select all

Sun Mar 15 07:23:11 AEDT 2020: executing enigma2
Sun Mar 15 08:14:14 AEDT 2020: enigma2 terminated with code 1
Sun Mar 15 08:24:42 AEDT 2020: executing enigma2

I did notice that in that log. 8:24:42 would have been when we turned it on from deep standby, & it remained either on while we were using it or in standby.

… when the "gone to lunch" sign was put on the door.

That would have been the ‘Gone fishing’ sign.

[MrQuade]

So will ‘init 3’ do that? I did think of it but wasn’t sure.

I'd do an init 4, then an init 3.

Or simply "reboot" to start from scratch. That's probably the "safest" in this instance, since the time to reboot is not really that important.
Edit:. Actually running a "shutdown" then a full or cycle would be absolutely safest if the hardware had been put into a funny state too.

[sub3R]

Yes, that would be the way to go in that situation.