Testers required for OpenViX Python 3 images

[twol]

I have BB_NUMBER_THREADS = "32" and it has always worked fine, but anyway that is nothing to do with compile threads.

I have 8 cores so it makes sense that I can run 8 compiles simultaneously... PARALLEL_MAKE = "-j 8", https://www.yoctoproject.org/docs/la...-PARALLEL_MAKE.

I run out of RAM just building the "nodejs" module on its own...

cd /home/openvix/5.6/builds/openvix/Py3/vuultimo4k
. env.source
bitbake nodejs

QT and nodejs will get rebuilt every time their repos are updated, which is pretty frequent.

BTW, so far no user has even noticed these are not in my builds.

The issue is those that use Stalker (QT) - that is the HiSilicon, Giga boxes at least would be impacted. Having said that I personally think Xstreamer is the better option, especially as there are (I believe) driver issues with Stalker.

[twol]

I have BB_NUMBER_THREADS = "32" and it has always worked fine, but anyway that is nothing to do with compile threads.

I have 8 cores so it makes sense that I can run 8 compiles simultaneously... PARALLEL_MAKE = "-j 8", https://www.yoctoproject.org/docs/la...-PARALLEL_MAKE.

I run out of RAM just building the "nodejs" module on its own...

cd /home/openvix/5.6/builds/openvix/Py3/vuultimo4k
. env.source
bitbake nodejs

QT and nodejs will get rebuilt every time their repos are updated, which is pretty frequent.

BTW, so far no user has even noticed these are not in my builds.

The issue is those that use Stalker (QT) - that is the HiSilicon, Giga boxes at least would be impacted. Having said that I personally think Xstreamer is the better option, especially as there are (I believe) driver issues with Stalker.

[birdman]

I have BB_NUMBER_THREADS = "32" and it has always worked fine, but anyway that is nothing to do with compile threads.

It has everything to do with the number of processes that will be running at once, since you're telling bitbake to run 32 builds in parallel. Each of these may also be running a build with as many processors as you actually have. And each process uses additional memory.

I have 8 cores so it makes sense that I can run 8 compiles simultaneously... PARALLEL_MAKE = "-j 8",

So each of the 32-parallel builds can use 8 processes at once. That potentially 256 concurrent compilations on an 8 core system. Likely to cause a lot of cache and TLB misses, slowing things down overall even before paging gets involved.

[Huevos]

It has everything to do with the number of processes that will be running at once, since you're telling bitbake to run 32 builds in parallel. Each of these may also be running a build with as many processors as you actually have. And each process uses additional memory.
So each of the 32-parallel builds can use 8 processes at once. That potentially 256 concurrent compilations on an 8 core system. Likely to cause a lot of cache and TLB misses, slowing things down overall even before paging gets involved.

No, PARALLEL_MAKE is the maximum number of concurrent compiles, i.e. 8, not 32 x 8. Anyway it runs out of resources just building the single module.

[BrokenUnusableAccount]

It has everything to do with the number of processes that will be running at once, since you're telling bitbake to run 32 builds in parallel. Each of these may also be running a build with as many processors as you actually have. And each process uses additional memory.
So each of the 32-parallel builds can use 8 processes at once. That potentially 256 concurrent compilations on an 8 core system. Likely to cause a lot of cache and TLB misses, slowing things down overall even before paging gets involved.

No, PARALLEL_MAKE is the maximum number of concurrent compiles, i.e. 8, not 32 x 8. Anyway it runs out of resources just building the single module.

I don't know exactly what happens but I always used

BB_NUMBER_THREADS = "4"
PARALLEL_MAKE = "-j 4"

And doing

ps -FHUopenvixbuilder

sometimes showed much more than 4 simultaneous compiles running and sometimes multiple instances of:

make -j 4

and/or

ninja -j 4

on the go at the same time.

[birdman]

No, PARALLEL_MAKE is the maximum number of concurrent compiles, i.e. 8, not 32 x 8. Anyway it runs out of resources just building the single module.

PARALLEL_MAKE is the option sent to make.
This means that each make command can fire off 8 sub-processes.
Separate make commands have no inter-process connexion, so if you have 2 they may use 16 sub-processes. etc.

I've not seen it run of of memory building node.js for python2 - and since it isn't a python build can't see why that should change for python3.

[birdman]

I don't know exactly what happens but I always used

BB_NUMBER_THREADS = "4"
PARALLEL_MAKE = "-j 4"

Whereas I delete these options before a build start so that the build uses the number of cores on the system that is doing the build for each option.
My build is on an external disk and may be run on systems with either 4 or 8 cores.

[thecaretaker]

@Caretaker .. Sounds like you have a crash - so look for crash log and also current & last debug log - post all !

Also try this version (which runs fine for me both direct to TV an also through a Onkyo amp)
Putty: init 4 (space between)
Filezilla: copy attachment to /usr/lib/enigma2/python/Components
Putty: init 6 (space between)

#########Note - attachment changed##### 16.39 19/7/2021

Using 5.5.14.008 now and I have the same problem. This time, your fix doesn't work (although I see the attachment has changed).

The box doesn't go into standby.

Debug Log: Enigma2_debug_2021-08-02_01-17-28.log

No crash log created.

[twol]

Using 5.5.14.008 now and I have the same problem. This time, your fix doesn't work (although I see the attachment has changed).

The box doesn't go into standby.

Debug Log: Enigma2_debug_2021-08-02_01-17-28.log

No crash log created.

can you try this and post log.

[Joe_90]

Is anyone having issues with the VPS plugin? It has been working perfectly for me for months, but several timers failed to record recently. When VPS starts monitoring the running status the normal log entry would be:

<  4278.6534> 17:18:01.2864 [RecordTimer] [VPS] start monitoring running-status
<  4278.6536> 17:18:01.2866 [eConsoleAppContainer] Starting /bin/sh
<  4278.6545> 17:18:01.2876 [VPS-Plugin] next execution in 30 sec
<  4278.7404> 17:18:01.3734 [eDVBDemux] open demux /dev/dvb/adapter0/demux0
<  4278.7405> 17:18:01.3735 [eDVBSectionReader] DMX_SET_FILTER pid=259
<  4278.9117> 17:18:01.5447 [eDVBDemux] open demux /dev/dvb/adapter0/demux0
<  4278.9119> 17:18:01.5449 [eDVBSectionReader] DMX_SET_FILTER pid=264
<  4279.0749> 17:18:01.7080 [eDVBDemux] open demux /dev/dvb/adapter0/demux0
<  4279.0752> 17:18:01.7082 [eDVBSectionReader] DMX_SET_FILTER pid=262
<  4279.2131> 17:18:01.8461 [eDVBServiceRecord] now running: †Flog It!‡ (1800 seconds)
<  4279.2132> 17:18:01.8462 [eDVBDemux] open demux /dev/dvb/adapter0/demux0
<  4279.2133> 17:18:01.8463 [eDVBSectionReader] DMX_SET_FILTER pid=18
<  4279.2146> 17:18:01.8476 [RecordTimer] [VPS] RUNNING_STATUS 1 FOLLOWING
<  4279.2151> 17:18:01.8481 [RecordTimer] [VPS] enable autoincrease

(This is from my GB Quad which is running DEV 5.4.014.008)

My AX61 HD is running Dev 5.5.014.008 and VPS seems to be failing to get a running status properly on satellite. The status coming up in the log is "EVENT_ENDED", even though the scheduled event has not begun yet. This has been happening for the past 10 days or so, but I can't give an exact timeframe as I don't record that much on the AX61. I don't recall any recent changes to VPS and I think that DEV 5.4 is pretty up-to-date with any commits made on the 5.5 py3 changes. The attached log snip shows a timer waiting to record "Room to Improve" on CH5 HD at 11.15am.

<  1871.4830> 11:14:24.2430 [RecordTimer] [VPS] start monitoring running-status
<  1871.4832> 11:14:24.2432 [eConsoleAppContainer] Starting /bin/sh
<  1871.4838> 11:14:24.2438 [VPS-Plugin] next execution in 30 sec
<  1871.5844> 11:14:24.3444 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1871.5846> 11:14:24.3446 [eDVBSectionReader] DMX_SET_FILTER pid=261
<  1871.7904> 11:14:24.5504 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1871.7905> 11:14:24.5506 [eDVBSectionReader] DMX_SET_FILTER pid=256
<  1871.9707> 11:14:24.7308 [eDVBServiceRecord] now running: †Jeremy Vine‡ (7200 seconds)
<  1871.9708> 11:14:24.7309 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1871.9709> 11:14:24.7309 [eDVBSectionReader] DMX_SET_FILTER pid=18
<  1871.9715> 11:14:24.7315 [RecordTimer] [VPS] EVENT_ENDED
<  1871.9716> 11:14:24.7316 [eConsoleAppContainer] user kill(SIGKILL)
<  1871.9717> 11:14:24.7317 [RecordTimer] [VPS] stop monitoring
<  1871.9717> 11:14:24.7318 [eDVBServiceRecord] stop recording!
<  1871.9718> 11:14:24.7318 [eDVBServiceRecord] (was not recording)
<  1871.9721> 11:14:24.7321 [eDVBCAService] free slot 0 demux 1 for service 1:0:19:1E46:809:2:11A0000:0:0:0:
<  1871.9721> 11:14:24.7321 [eDVBCAService] free service 1:0:19:1E46:809:2:11A0000:0:0:0:
<  1871.9728> 11:14:24.7328 [eDVBResourceManager] start release channel timer
<  1871.9729> 11:14:24.7329 [RecordTimer] [VPS] stop RecordService (simulation)
<  1871.9731> 11:14:24.7331 [RecordTimer] [VPS] stop monitoring (process terminated)
<  1872.6690> 11:14:25.4291 [eEPGChannelData] start reading events(1627899265)
<  1872.6692> 11:14:25.4292 [eDVBSectionReader] DMX_SET_FILTER pid=3842
<  1872.6695> 11:14:25.4295 [eDVBSectionReader] DMX_SET_FILTER pid=3003
<  1872.6696> 11:14:25.4296 [eDVBSectionReader] DMX_SET_FILTER pid=18
<  1872.6698> 11:14:25.4298 [eDVBSectionReader] DMX_SET_FILTER pid=18
<  1872.6700> 11:14:25.4300 [eDVBSectionReader] DMX_SET_FILTER pid=18
<  1872.6702> 11:14:25.4302 [huffman] read.. '/usr/share/enigma2/otv_011a0000_0002_0809.dict'

<snip>

<  1933.5838> 11:15:26.3438 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1933.5838> 11:15:26.3438 [eDVBSectionReader] DMX_SET_FILTER pid=0
<  1933.5843> 11:15:26.3443 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1933.5843> 11:15:26.3443 [eDVBSectionReader] DMX_SET_FILTER pid=267
<  1933.5844> 11:15:26.3444 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1933.5845> 11:15:26.3445 [eDVBSectionReader] DMX_SET_FILTER pid=17
<  1933.7820> 11:15:26.5420 [eDVBFrontend0] set dynamic current limiting
<  1933.7843> 11:15:26.5443 [eDVBServiceRecord] RECORD service event 5
<  1933.7844> 11:15:26.5444 [eDVBServiceRecord] Recording to ...
<  1933.7844> 11:15:26.5444 [eDVBServiceRecord] can't open recording file: No such file or directory
<  1933.7850> 11:15:26.5450 [eDVBCIInterfaces] gotPMT
<  1933.7850> 11:15:26.5450 [eDVBCAService] don't build/send the same CA PMT twice
<  1933.7851> 11:15:26.5451 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1933.7852> 11:15:26.5452 [eDVBSectionReader] DMX_SET_FILTER pid=257
<  1933.7904> 11:15:26.5504 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1933.7905> 11:15:26.5505 [eDVBSectionReader] DMX_SET_FILTER pid=258
<  1933.9884> 11:15:26.7484 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1933.9885> 11:15:26.7485 [eDVBSectionReader] DMX_SET_FILTER pid=259
<  1934.1875> 11:15:26.9475 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1934.1877> 11:15:26.9477 [eDVBSectionReader] DMX_SET_FILTER pid=260
<  1934.2362> 11:15:26.9963 [RecordTimer] [VPS] start monitoring running-status
<  1934.2364> 11:15:26.9964 [eConsoleAppContainer] Starting /bin/sh
<  1934.2374> 11:15:26.9975 [VPS-Plugin] next execution in 30 sec
<  1934.3855> 11:15:27.1455 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1934.3857> 11:15:27.1457 [eDVBSectionReader] DMX_SET_FILTER pid=261
<  1934.5915> 11:15:27.3515 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1934.5917> 11:15:27.3517 [eDVBSectionReader] DMX_SET_FILTER pid=256
<  1934.6318> 11:15:27.3918 [eDVBServiceRecord] now running: †New: Room To Improve‡ (3900 seconds)
<  1934.6319> 11:15:27.3919 [eDVBDemux] open demux /dev/dvb/adapter0/demux1
<  1934.6322> 11:15:27.3923 [eDVBSectionReader] DMX_SET_FILTER pid=18
<  1934.6329> 11:15:27.3929 [RecordTimer] [VPS] EVENT_ENDED
<  1934.6330> 11:15:27.3930 [eConsoleAppContainer] user kill(SIGKILL)
<  1934.6332> 11:15:27.3932 [RecordTimer] [VPS] stop monitoring
<  1934.6332> 11:15:27.3933 [eDVBServiceRecord] stop recording!
<  1934.6333> 11:15:27.3933 [eDVBServiceRecord] (was not recording)
<  1934.6336> 11:15:27.3937 [eDVBCAService] free slot 0 demux 1 for service 1:0:19:1E46:809:2:11A0000:0:0:0:
<  1934.6337> 11:15:27.3937 [eDVBCAService] free service 1:0:19:1E46:809:2:11A0000:0:0:0:
<  1934.6347> 11:15:27.3947 [eDVBResourceManager] start release channel timer
<  1934.6349> 11:15:27.3949 [RecordTimer] [VPS] stop RecordService (simulation)
<  1934.6351> 11:15:27.3951 [RecordTimer] [VPS] stop monitoring (process terminated)
<  1935.4126> 11:15:28.1726 [eEPGChannelData] start reading events(1627899328)
<  1935.4127> 11:15:28.1727 [eDVBSectionReader] DMX_SET_FILTER pid=3842
<  1935.4129> 11:15:28.1729 [eDVBSectionReader] DMX_SET_FILTER pid=3003
<  1935.4131> 11:15:28.1731 [eDVBSectionReader] DMX_SET_FILTER pid=18
<  1935.4132> 11:15:28.1732 [eDVBSectionReader] DMX_SET_FILTER pid=18
<  1935.4134> 11:15:28.1734 [eDVBSectionReader] DMX_SET_FILTER pid=18

The event change from "Jeremy Vine" to "Room To Improve" is shown in the log, but the running-status is consistently shown as "EVENT_ENDED" in VPS and the programme is not recorded consequently. Strangely, VPS is working fine on the Irish terrestrial service Saorview on the same box.

Any suggestions? I was going to download and flash a clean 5.5 image, but the Image Manager is only showing me 5.4 images.

[johnhoe]

http://images.satellites-xml.org/Py3/

[thecaretaker]

can you try this and post log.

No joy, the VU+ front screen goes dim and shows the date and time. TV picture turns off. Then the VU+ screen goes full brightness and back to the Picon and TV shows channel again. I tried it several times. Hopefully you'll be able to see that in the log.

Enigma2_debug_2021-08-02_14-39-13.log

[Joe_90]

http://images.satellites-xml.org/Py3/

@twol - do you have the zip images for the ax61 hosted? The ax61 zip in @Huevos's repository above is several versions old. When I follow the URL for your builds all I can find are the update files, not the actual image zips.

[ccs]

...you should be able to see 5.5/Py3 images in imagemanager.

[twol]

@ccs is correct - and I can see the ax hd51 latest in ImageManager under ViX

If that is not working, we need to find out why not, but I have moved the latest (008) to my openvix-builds folder