Register of Weird Crash Symptoms

[Black Hole]

Thought perhaps we should gather crash symptoms in one place.

 

[Black Hole]

I used WebIF to access the "headless" HDR-FOX I use to make recordings from radio and convert for later consumption in the car. Some scheduled recordings were missing. The WebIF schedule browser showed no active EPG entries for them (when they should have been rescheduled for next week as continuing series).

Foolishly I ran fix-flash-packages and thereby killed ir (there is a problem with the ir package at the moment which means it can't be reinstalled when fix-flash-packages deletes it and tries to reinstate it).

Rebooting fixed it: a crash must have stopped settop but not WebIF, so the EPG wasn't being updated and recordings were not being made, despite still having access to WebIF.

 

[/df]

...
Rebooting fixed it: a crash must have stopped settop but not WebIF, so the EPG wasn't being updated and recordings were not being made.

One hopes that it must have hung as the system should have rebooted if settop crashes.

My most common weirdness ...

Video partition (at least) gets marked as RO, settop possibly hangs (audio can continue, video freezes, no response to remote or buttons). On crashing or being killed or s/w reboot forced, HDMI doesn't work. Power cycling restores it.

 

[Black Hole]

One hopes that it must have hung

I dare say; I made no distinction.

 

[Black Hole]

The same "headless" HDR-FOX as mentioned perviously: it's on WiFi and I couldn't get WebIF access even though the VFD was scrolling as usual. However, it was not responding to handset or front panel inputs, and the TV just gave a green screen.

A power cycle restored normal function.

 

[Black Hole]

I was watching a recording, and went into the menu system to check a detail, but couldn't exit. The HDR-FOX was stuck with the menu on-screen, but the recording continuing to play in the background. No response to handset or front panel controls, but WebIF working and VFD scrolling. No response to WebIF virtual remote.

After a power-cycle all back to normal, but there was no resume offered on the recording.

 

[prpr]

I've had apparent IR lockups several times as well. Only way out is to power cycle, which is a nuisance if it's recording of course.

 

[sine24]

It seems to have happened twice recently: playng back a recording the t2 crashed, no webif. I've had crashes before but don't remember one during playback.

 

[Black Hole]

This one really has me scratching my head. It's not exactly a "crash" or a "freeze", and I was ready to declare it a hardware failure (my "HDR3"):

"No programmes are currently being broadcast on this channel" (despite that not being correct, and signal quality showing 100%). The strange thing was, if I changed service (any service, any mux), I got programme for maybe 20 seconds and then back to black. Signal detection showed OK on all muxes.

Tuner? I set it recording and then switched to another mux - same thing. Tuner test in the service menu showed OK on all muxes on both tuners.

Heat? Exhaust was warm, but the fan was running and Sysmon showed a steady 43deg.C. Maybe it needs dusting out? Not easy - I have the (captive) mains lead cable-tied with other leads (a mains inlet socket would be a useful mod after all!).

So, just for the hell of it I tried a (quick and easy) power cycle... not long enough to have alterered the temperature, but in the hour since the problem has not recurred. All recordings made since (and including) 7.55am on Thursday 18th Feb are broken (0 mins), the previous good recording was completed on 7.04pm Wednesday.

What could cause this? The only thing which occurs to me is that the decoder needs periodical commands to keep it alive (can't imagine why), and the individual process which sends those periodic refreshes had crashed (leaving everything else alive). We know that if settop crashes, at least the audio stream continues to be delivered to the HDMI (which was not the case here), so maybe the continuation of audio requires settop to have crashed so it can't notice something wrong and put up an on-screen message.

 

[Black Hole]

HDR1: Following a WebIF reboot, WebIF working but the "ticker" shows "Cannot find humaxtv process":


...and ditto on the WebIF >> Diagnostics >> Reboot screen:


SMB server working (can play HDR1 recordings on other machines). Live picture frozen. Front panel shows "Reboot...":


Reboot from WebIF >> Diagnostics does not fix it, I suspect nothing actually happens. Power-cycle fixed it.

I had been fiddling with /mod/etc/init.d, hence had performed a WebIF reboot. Live TV froze at that time (Red Button), so I think the reboot failed and got stuck without actually rebooting.

There has been a lot of alteration lately with beta firmware and beta WebIF, but I have been unable to reproduce this on my other machines with the same firmware and WebIF. I have also been unable to reproduce this on HDR1, despite recreating my init.d fiddling.

 

[/df]

So, fiddling with some configuration database stuff at boot-up (xinit), one can end up with a 4:3 letterbox display on a blue background with country set to Australia and the channel-toggle function of Back on the I-plate disabled. Then setting Preferences>Video>Screen Ratio causes an actual CrASh. Factory Default restores the expected functionality.

This is a comparison of TBL_MENUCONFIG from setup.db for the bad and good cases (the same comparison for TBL_USERCONFIG is uninteresting):

attach '/var/lib/humaxtv/setup.db' as su1;
attach '/mod/tmp/bak/humaxtv/setup.db' as su2;
select weird.*, good.itemValue, good.itemText from su2.TBL_MENUCONFIG weird, su1.TBL_MENUCONFIG good 
     where weird.itemName = good.itemName and 
                weird.itemValue||weird.itemText <> 
                     (select itemValue||itemText from su1.TBL_MENUCONFIG where itemName = weird.itemName);

4|MENU_LANG|0||(null)|11|
6|SOUND_DIGITALOUTPUT|0||(null)|3|
10|RESOLUTION|12||(null)|15|
12|EMPTY_BOX_COLOR|0||(null)|1|
15|AUTOMATIC_POWER_DOWN|0||(null)|1|
16|WLAN_IS_CONNECTED|1||(null)|0|
19|AUDIO_LANG|0||(null)|11|
28|INFO_DISPLAY_TIME|0||(null)|10000|
29|OSD_TRANSPARENCY|0||(null)|1|
37|FAV_CUSTOM_STR01|1|Faves|(null)|0|Favourite 1
38|FAV_CUSTOM_STR02|1|FavesHD|(null)|0|Favourite 2
41|FAV_CUSTOM_STR04|1|Favourite 4|(null)|0|Favourite 4
44|SUBTITLE_LANG|0||(null)|11|
48|POWER_ON_CHANNEL|131098||(null)|-1|
49|TIME_VOLUME|10||(null)|0|

For items whose name includes LANG, 0 must be Australia (first, or zeroth, in the list of countries in the hidden menu's Country pull-down) and 11 must be England.

EMPTY_BOX_COLOR seems to determine whether unpainted (no signal) areas are blue (0) or black (1), but doesn't have a corresponding UI, I think.

RESOLUTION presumably maps to V.Fmt.

TIME_VOLUME is a mystery, though it is known to the settop blob.

The breakage of channel toggle also eludes explanation.

After further investigation, apparently this is what happens when you accidentally set all itemValues to 0 in TBL_MENUCONFIG.

 

[/df]

HDR1: Following a WebIF reboot, WebIF working but the "ticker" shows "Cannot find humaxtv process":
...
I had been fiddling with /mod/etc/init.d, hence had performed a WebIF reboot. Live TV froze at that time (Red Button), so I think the reboot failed and got stuck without actually rebooting.

Apparently "Cannot find humaxtv process" was an accurate diagnostic. Presumably rebooting or reinstalling made everything work again.

 

[bottletop]

Was watching live TV on HDR. Nodded off and woke up to blank screen, no sound. No WebIF and pinging the ip didn't work. Switched HDMI sources and back again on TV, still blank. Didn't respond to any IR handset functions. Had to power off and on the HDR.

Edit: Although there isn't a message, because it's just a blank screen, the TV reports something like "unsupported mode" or "unsupported resolution".

 

[MymsMan]

A runaway recording that wasn't!

Not a crash as such but required reboot to resolve

Yesterday I had what appeared to be a runaway recording, It was listed as an active recording in the webif long after it should have completed and subsequent recordings failed to start recording,
It couldn't be stopped by the remote but when I rebooted the machine I found the recording was actually the shorter by about 5 minutes than the correct length and was truncated rather than runaway

 

[prpr]

Last night's weirdness:

humax /mnt/hd2/My Video/Olympics 2020 # ls -lrt *.ts
-rw-r--r-- 1 root root 10524938240 Fri  6 Aug 2021 03:08:01 Olympics 2020 Live_20210806_0005.ts
-rw-r--r-- 1 root root           0 Fri  6 Aug 2021 04:30:36 Olympics 2020 Live_20210806_0430.ts
-rw-r--r-- 1 root root           0 Fri  6 Aug 2021 05:45:21 Olympic Breakfast_20210806_0545.ts
-rw-r--r-- 1 root root           0 Fri  6 Aug 2021 09:02:54 Olympics 2020 Live_20210806_0902.ts

humax /mnt/hd2/My Video/Olympics 2020 # ls -lrt *.nts
-rw-r--r-- 1 root root 28193024 Fri  6 Aug 2021 03:08:05 Olympics 2020 Live_20210806_0005.nts
-rw-r--r-- 1 root root 11564800 Fri  6 Aug 2021 05:45:15 Olympics 2020 Live_20210806_0430.nts
-rw-r--r-- 1 root root 27495296 Fri  6 Aug 2021 09:02:52 Olympic Breakfast_20210806_0545.nts
-rw-r--r-- 1 root root  9607872 Fri  6 Aug 2021 10:01:49 Olympics 2020 Live_20210806_0902.nts

The .ts files stopped recording at 03:08 and nothing recorded after that, BUT, all the .nts files are apparently correct, as are the .hmt files, so it looks like something just stopped the elemental streams being pushed to disk (note the timestamps between .ts and .nts).
Live pictures and audio were still working, as was all the CF/WebIf stuff.
Attempting to power off with the remote did nothing, so it was time for the Big Red Switch.

 

[Black Hole]

That is really weird!

 

[MymsMan]

The .ts files stopped recording at 03:08 and nothing recorded after that, BUT, all the .nts files are apparently correct, as are the .hmt files, so it looks like something just stopped the elemental streams being pushed to disk (note the timestamps between .ts and .nts).
Live pictures and audio were still working, as was all the CF/WebIf stuff.
Attempting to power off with the remote did nothing, so it was time for the Big Red Switch.

I have seen it a few times in the past - it is one of the reasons I put a check into schedchk to see if the .ts files for supposedly active recordings were actually growing

 

[Black Hole]

I have seen it a few times in the past - it is one of the reasons I put a check into schedchk to see if the .ts files for supposedly active recordings were actually growing

Good idea, but until the box is rebooted...

I really must get around to implementing my nightly timered brief power cut (within a power-off schedule), but mechanical timeswitches can drift (especially if there is a real power cut). Maybe Gosunds...

 

[prpr]

I really must get around to implementing my nightly timered brief power cut

Probably good enough with just a software reboot.

Maybe Gosunds...

They seem to show you foreign products and prices, even when you select the UK option. Not impressed.

 

[Black Hole]

Probably good enough with just a software reboot.

For that, maybe, but a software-invoked reset wouldn't overcome a more serious freeze.

They seem to show you foreign products and prices, even when you select the UK option. Not impressed.

I bought some on an Amazon discount, and actually they work really well. I would need some more though.