Friday, 2017-07-21

*** tpb has joined #timvideos00:00
CarlFKnote to anyone testing: installing to ssd  over and over is super nice00:07
olasdexcept for the ssd00:09
olasdbut yes, SSD and installer on USB 3 stick = <3 <300:09
CarlFKyeah, I got a usb3 stick just for this.  and now I am installing to a bunch of spinny disks00:14
CarlFKI wonder if I can boot off an sd card...00:16
*** rohitksingh has joined #timvideos01:07
Ishan_Bansalshenki : ping01:11
*** CarlFK has quit IRC01:34
*** rohitksingh has quit IRC02:04
*** sb0 has quit IRC02:29
*** rohitksingh_work has joined #timvideos03:33
mithroAfternoon people03:54
mithroOffice hours will only be 30 minutes today03:55
mithroIshan_Bansal: Do you need any help?03:55
mithropaddatrapper: How goes things?03:55
*** CarlFK has joined #timvideos04:03
*** ChanServ sets mode: +v CarlFK04:03
mithroxfxf: I do think we should get into the situation where we are reimaging machine on a very regular bases04:04
xfxfmithro: sure, but the ansible infrastructure built during LCA seems broken now04:06
mithroxfxf: I feel like olasd / jea are having a different experience, so it would be good to figure out why -- totally understand that you're short on time with PyCon rapidly approaching....04:07
xfxfi want to figure out why too, but pycon prep is only me, with thrlr additional fun of no access to the rooms until 5 am the morning of the event04:08
mithroxfxf: Gah :-(04:08
xfxfhence using an install i know works and can be placed in. we have literally zero time to debug anything new that comes up04:08
mithroxfxf: Yeah04:08
xfxfbut agreed, will reimage after pycon04:09
CarlFKmithro: symlink04:12
mithroCarlFK: not happening today, remind me again tomorrow04:12
mithroCarlFK: but do keep poking me, eventually I'll get annoyed enough to do it :-P04:12
*** SamSagaZ has quit IRC04:23
*** SamSagaZ has joined #timvideos04:24
paddatrappermithro: yeah, all good. Need to work out why I'm getting EOVERFLOW as a response to every iso request07:59
mithropaddatrapper: Do you need me to merge / look at stuff?08:00
paddatrappermithro: the only thing outstanding is the audio descriptors. I've branched off it for what I'm currently doing, so that can be merged. Otherwise I can close it and open a new one with the latest stuff08:02
tpbTitle: Audio descriptors by paddatrapper · Pull Request #23 · timvideos/HDMI2USB-fx2-firmware · GitHub (at
mithropaddatrapper: So how do I replicate your work?08:28
mithropaddatrapper: Are you going to setup travis?08:28
paddatrappermithro: replicate which work?08:31
mithropaddatrapper: The audio FX208:31
paddatrappermithro: yeah I intend to this weekend08:31
paddatrappermithro: checkout audio-data from my fork. cd audio && make08:32
paddatrapperThat will give you the latest of my work there08:33
mithroHow do I load it?08:34
paddatrappermake load08:34
paddatrapperThat will use hdmi2usb-mode-switch to load the audio.hex08:34
mithropaddatrapper: BTW You might want to look at the sigrok-firmware-fx2lafw sometime...09:05
paddatrappermithro: I use it regularly?09:16
paddatrapperDo you mean look at how it is written?09:16
mithropaddatrapper: Yeah09:24
paddatrappermithro: ok I'll take a look09:32
mithroFirst FOSS usage of the gpif in the FX2 I've seen ->;a=blob;f=gpif-acquisition.c;h=7d3dcb6d59d8c1c69048ce7b3efbca98d80e5d21;hb=HEAD09:56
tpbTitle: Git - sigrok-firmware-fx2lafw.git/blob - gpif-acquisition.c (at
mithropaddatrapper: We should diff fx2lib to the fx2lib in sigrok-firmware09:57
paddatrappermithro: that's a good idea09:57
paddatrappermithro: it's good to see that they don't know what's happening in some of their code either09:59
mithropaddatrapper: Oh? :-P09:59
paddatrappermithro: scanning through I spotted a couple of question marks :)10:00
mithropaddatrapper: Got examples?10:22
paddatrappermithro: seems it is only line 45. Though there were more..10:32
jeamithro: building litex firmware for opsis - I do want to set platform=opsis, target=hdmi2usb and then run `make gateware`?10:49
mithrojea: Yes - platform/target in caps10:49
jeait is building at the moment. just taking a long time10:49
jeawhich i expect, but just didn't want to get all the way through and then find I did the wrong thing10:50
mithroIt will take about 15-30 minutes10:50
jeathis one has been routing for 57+ minutes10:50
jeashould I stop and start again?10:51
mithroIf it takes longer than 30+ minutes then it will fail10:51
jeasounds like i should do this again then10:51
mithrojea: Assuming your computer is similar performance as mine10:51
jeathis is on a gen 7 core i5 @ 2.2ghz10:52
jeamithro: what is the difference between video and hdmi2usb targets?10:55
mithrojea: video doesn't have capture functionality (IE no jpeg encoder)10:55
jeaah right10:55
mithroThe video target should be around 5-10 minutes11:03
jeamithro: is there a hdmi2usb prebuilt bin for the opsis?11:05
jeaI see one for atlys, but not opsis, unless i am in the wrong place11:05
tpbTitle: HDMI2USB-firmware-prebuilt/archive/master/v0.0.3-710-g947321b/opsis/hdmi2usb/lm32 at master · timvideos/HDMI2USB-firmware-prebuilt · GitHub (at
jeaah, just not in the latest build11:06
jeathanks mithro11:06
mithrojea: Yeah - that seems to be broken on travis - I'll try building here...11:07
mithroMaybe it really is broken11:10
jeaoh ok11:18
jeathat isn't too good11:18
jeawe are going to use the prebuilt one and see how it goes11:18
mithrojea: Normally it just push a new change and it starts building again - stupid non-determinisim11:19
jeamithro: that is fpgas for you11:20
mithrojea: Well more the horrible ISE software...11:21
mithrojea: Hrm, I'm getting the same behaviour here....11:27
jeamithro: I'm thinking of finding a commit from around LCA and then building that11:28
mithrojea: Have you tried the one I just posted?11:28
jeawhere did you post that? in the prebuilt repo?11:28
mithrojea> ah, just not in the latest build11:28
jeaoh that one11:29
jeavys just flashed with it, but doesn't think it worked11:29
vysError: xc6s.tap: IR capture error; saw 0x03 not 0x0111:30
mithrovys: Try flashing again....11:30
vysthat's from the #710 build11:30
vysthe DNA was 0x18181818181818....11:30
vysI assume that's incorrect11:30
vysnow it's decided that it's 0x1381c...11:30
mithroSee the docs11:30
vysI think we're ok now11:31
tpbTitle: HDMI2USB-litex-firmware/ at master · timvideos/HDMI2USB-litex-firmware · GitHub (at
mithroBTW Did I ever share ?11:45
vysseem to be unable to load firmware onto the opsis11:51
vyshow long should it normally take?11:52
mithrovys: about 1-2 minutes11:53
mithrovys: Don't forget to power cycle it afterwards11:53
vysloading or flashing11:55
vysI presume the firmware relies on the gateware having a particular version?11:56
mithrovys: You just need to flash the one .bin file11:56
mithroThe flash.bin11:57
mithrovys: That contains the gateware+bios+firmware11:57
vysoh right11:58
vyswhich opsis-mode-switch command for that?11:58
mithrovys: You use the --flash-gateware command11:59
vysI'll try11:59
mithroNeed to add a --flash-image command to the firmware11:59
vysI do, or you do?11:59
mithroIf I say you do, will you do it for me? :-P12:00
vysnot tonight12:00
vysbut potentially12:00
mithroI'm not doing it tonight12:00
vysthe opsis specifies `--flash-gateware=$(TARGET_BUILD_DIR)/gateware/top.bin`12:00
mithrovys: It would be really nice if the hdmi2usb-mode-switch had a "--download-gateware=stable" that would pull from the prebuilt too12:01
vysso I use opsis-mode-switch with --flash-gateware, and specify the flash.bin from the prebuilt archive?12:02
vysand then power cycle it, and it should all come good?12:02
mithroIn _theory_12:03
mithrovys: I thought I had figured out how to make the opsis reset after programming, but it didn't seem to be working last I tested it12:05
mithroI'm about to head out....12:07
vysI'm having trouble flashing it, but will continue12:08
vysmithro: thanks for your help tonight12:08
mithropaddatrapper: Apparently the FX2 miniboard is called a "Lcsof CY7C68013A Mini Board" ->
tpbTitle: Lcsoft Mini Board - sigrok (at
*** sb0 has joined #timvideos12:29
*** rohitksingh_work has quit IRC12:36
*** sb0 has quit IRC12:45
mithrovys: any luck?13:03
mithrojea: ^13:05
vysI'm now flashing firmware reasonably well13:08
vysalthough running 'm l' causes the whole opsis to lock up on the 710 build13:08
mithrohrm.... 'm l'?13:09
vysvideo_mode list13:10
mithrovys: It crashes with both? or just the short one?13:12
*** sb0 has joined #timvideos13:19
vyspotentially just hangs for a really long time13:19
mithroWhat _does_ work?13:20
vysI can connect one of our cameras to an encoder13:22
vysand occasionally cheese will display it to me13:22
vyswhen cheese doesn't hang indefinitely13:22
vysrunning 'm l' pauses the clock on the encoder13:23
CarlFKvys: you have 2 opsis boards right?13:26
vysCarlFK: ^13:26
CarlFKlol - k.  if only 1 I would dig mine up to try on different hardware13:27
vysmithro: setting it to mode 14 dies13:27
CarlFKyou are using this build?   (06:05:58 AM) mithro: jea:
tpbTitle: HDMI2USB-firmware-prebuilt/archive/master/v0.0.3-710-g947321b/opsis/hdmi2usb/lm32 at master · timvideos/HDMI2USB-firmware-prebuilt · GitHub (at
mithroCan you give me the output of status?13:30
vyssetting it to mode 14 kills the clock on the pattern page too13:30
vysH2U 00:00:21>s13:30
vysinput0:  0x0 (@47.94 MHz) (capturing)13:30
vysinput1:  0x0 (@0.00 MHz) (capturing)13:30
vysoutput0: [email protected] from input1 (underflows: 0)13:30
vysoutput1: [email protected] from input1 (underflows: 0)13:30
vysEDID primary mode:   [email protected]13:30
vysEDID secondary mode: [email protected]13:30
vysencoder: 1280x720 @ 0fps from input1 (q: 85)13:30
mithroTurn off output1 and input113:30
vyswith them off 'm l' still dies13:32
vysI should just flash older firmware13:32
mithropaddatrapper: This page lists a lot of the fx2 based devices -
tpbTitle: fx2lafw - sigrok (at
jeaCarlFK: using the latest ansible stuff, should a mix computer do an auto-login and then open up voctogui?13:47
jeabecause i am seeing it do an auto-login, but nothing opens. and opening voctogui from the applications menu freezes the whole machine13:48
CarlFKjea: auto log in yes, but not run the gui - click the cat in the toolbar13:54
mithrojea: That sounds like the voctogui or core isn't using the correct config?13:54
CarlFKfreezes the whole machine - i haven't seen that13:55
mithroCarlFK: I have when you start it up with 1080p13:55
jeaCarlFK: I don't have a cat in the toolbar13:56
jeaso I'm not sure what went wrong there13:56
jeamaybe i stuffed up the inventory13:57
tpbTitle: joeladdison / wordcamp-bne-2017 Bitbucket (at
CarlFKjea: how latests is latest? ;)  like do you have a cfg file that looks like this:
tpbTitle: Ubuntu Pastebin (at
jeawell it was installed last night / tonight13:59
jeaso pretty latest14:00
tpbTitle: joeladdison / wordcamp-bne-2017 / source / inventory / hosts Bitbucket (at
CarlFKthat looks ok14:02
CarlFKjea: how did you install the OS?14:04
jeaCarlFK: I used the latest CD ISO, then installed using the same preseed config as in your repo14:05
CarlFKjea: debian or ubuntu?14:06
jeaafter that, I ran ansible to set everything up, as per late-command.sh14:06
CarlFKwhats the iso filename?14:06
CarlFKtrying to figure out how far off it is from what we have done14:07
jeajust finding it now14:07
CarlFKthats a full desktop install right?  like you get gui and apps and everythng ?14:09
CarlFKIm guessing thats what broke it14:09
jeabut we installed the least amount possible, and followed the same settings as in the preseed14:09
jeait may well be what broke it. but now i need to fix it14:10
CarlFKleast amount  - like does that allow you to not install a dm ?14:11
jeai think there was an option for that14:11
CarlFK13 hours ago I setup 3 boxes from scratch -
tpbTitle: av-foss-stack/hosts at pyohio · xfxf/av-foss-stack · GitHub (at
CarlFKwhich is the same as yours14:14
CarlFKhmm, no opsis14:14
CarlFKthe 3 cnt boxes are 10 min drive away running test feeds to burn in the new CPUs I just swapped in..14:15
CarlFKbut I have the gator box here that I can mess with .. I'll see what happens if I add opsis to it14:16
jealooking in dmesg, the opsis is being continuously registered as a new uvcvideo thing14:20
jeawhich seems bad14:20
CarlFKjea: that is expected if it isn't plugged in or working14:20
jeawhat counts as working? it is definitely plugged in and on14:21
CarlFKwhich might be why things are 'bad'14:21
CarlFKshowing up as a uvc device14:21
jeaso if i have an opsis connected via USB, and it is turned on, it should keep being registered as a uvcvideo device14:22
jeain an interesting development, vys just ran voctogui under strace and it works14:23
CarlFKit should register once14:23
CarlFKtumbleweed: halp...14:24
CarlFKfrom memory (ill dig into code in a bit) ingest tries to send to vocto, when it exits it rmmod/modprobes the uvc driver to reset the opsis14:25
CarlFKor something like that14:25
jeaone difference would be that i enabled hdmi2usbd. maybe i should remove that14:25
CarlFKI don't use that14:25
jeai am very tempted to just use the last commit from lca and install everything again14:28
jeabut it sounds like you have the current master of your repo working14:28
CarlFKand tumbleweed reviewed and synced it with DC's repo even14:29
CarlFKhow do I join this #chan from matrix?14:30
jeasure, but that doesn't necessarily mean to me that it has been tested14:30
jea(i have seen people merge things without any review or testing at work, so i am sure it can happen elsewhere)14:30
CarlFKwe haven't changed what the end result is14:31
CarlFKwith the exception of the pxe server14:32
CarlFKbut the vocto and opsis stuff is all the saem14:32
CarlFKhow did you enable hdmi2usbd ?14:34
jeai used the compiled version from earlier this year14:37
jeai'll remove that and see what happens14:37
*** CarlFK[m] has joined #timvideos14:43
jeawelcome CarlFK[m] :)14:44
* CarlFK[m] uploaded an image: file1500648380200.jpg (22KB) <>14:47
CarlFK[m]ta da14:47
CarlFK[m]how close is that to yours?14:48
jeapretty similar, but no three buttons on the bottom toolbar at the end14:49
jeaon a grab machine, i did get the buttons appearing (but with two of them pointing at nothing, i think because grab has no vocto)14:49
CarlFK[m]jea: [email protected]:~$ stop-all-the-things14:57
CarlFK[m]that will shut down all vocto stuff14:57
CarlFK[m]voctocore -vvvv14:59
CarlFK[m]   DEBUG ConfigParser: successfully parsed config-files:14:59
CarlFK[m]DISPLAY=:0 voctogui -vvvv15:01
CarlFK[m](I'm sshed into gator15:01
CarlFK[m]also.. I have to change /etc/voctomix/voctogui.ini  ... use system=x15:03
jeaanother difference is mine is that it is using en_AU instead of en_US. I don't think that should matter, but it is something I had to comment out in the user role task (to set it to be en_US if it isn't already)15:10
* CarlFK[m] uploaded an image: file1500649818619.jpg (27KB) <>15:11
jeathat is definitely a voctogui15:12
CarlFK[m]voctomix-ingest ; voctomix-ingest --port 1000115:12
CarlFK[m](well, each in it15:12
CarlFK[m](well, each in its own shell)15:13
jeahow much disk space does a day of recordings take up?15:17
CarlFK[m]5 g per hour... DC will be 35G per day15:17
jeafrom memory we use 1tb drives at lca and that lasts 3 days at least15:17
CarlFK[m]hmm 35 is on the low end15:18
tpbTitle: veyepar - Show: Debian Debconf17 (at
CarlFK[m]I think those are using 10G / hour .. which was what dv used15:18
* jea[m] uploaded an image: ima_df6663d.jpeg (66KB) <>15:19
CarlFK[m]one 'problem' is all the space that gets used up testing.. and then we don't bother to clean it off before the show starts15:19
jeathat looks a lot better. now to see if voctogui opens15:19
CarlFK[m]because we are still testing when the show starts :p15:19
jeahaha yes15:19
CarlFKI see the vocto-cat !15:20
jeawell voctogui opens now. and the computer has not frozen15:22
jeaif I run the ingest things above, does that give me test sources?15:22
jeahmm. i'm not seeing pattern in vocto15:24
jeathe panels that should be showing video are just displaying the contents of the screen underneath the window15:24
* jea[m] uploaded an image: ima_5bfba3c.jpeg (112KB) <>15:25
CarlFK[m]also.. I have to change /etc/voctomix/voctogui.ini  ... use system=x15:28
CarlFK[m]do that15:28
CarlFK[m]or try the one in the middle:  #system=xv15:28
jeaok, let me try that15:28
jeasystem=x works15:31
jeanow to see if the opsis works15:31
jeahow do you normally connect to the opsis to send commands?15:34
jea(there is no flterm)15:34
CarlFK[m]which should default to all the right things... haven't tested that15:36
jeayep, it does15:37
jeaI'm not seeing any opsis output in vocto15:40
jeagrabber is the test pattern from voctomix-ingest15:40
jeai presume the camera should be coming from the opsis connected to the machine15:41
jea(also a bit concerning is the computer monitor going on and off)15:42
CarlFKlets start with the local one - opsis set to mode ... I forget. - whatever 720p is15:44
jea9 or 1015:45
tpbTitle: video-stack-deploy/videoteam-ingest.j2 at master · CarlFK/video-stack-deploy · GitHub (at
CarlFKvoctomix-ingest --video-source hdmi2usb15:48
CarlFKvoctomix-ingest --video-source hdmi2usb --video-attribs device=/dev/hdmi2usb/by-num/all0/video15:48
jeai can see that running15:55
CarlFK[m]in voctogui?15:58
CarlFK[m]what do you mean by "see that running" ?15:59
jeai did a `ps aux | grep vocto` and I see it there16:00
jeabut nothing coming into voctogui16:00
CarlFK[m]did you run it in a shell, or is that from the service thingy?16:01
jeaok. so if i stop all the things, then start all the things, should the ingest start too?16:01
jeathat was from the service thing16:01
jeai'll try running it again16:01
CarlFK[m]I love trying to guess what you are doing.  so much fun!16:02
CarlFK[m]stop all the things16:02
CarlFK[m]run each command in a shell16:02
jeasorry about that16:03
jeai have now run that command in the shell, and it stops with an error16:03
jeabroken pipe sending data16:04
CarlFK[m]I know this one!!!16:04
* jea waits with anticipation16:04
CarlFK[m]please please tell me you have a minicom or whatever term connection16:04
jeaah yep16:05
jealet me guess. i have to close it16:05
CarlFK[m]close that16:05
CarlFK[m]re-run ingest16:05
CarlFK[m]yeah.. that one gets me every time.16:05
jeawhat is minicom exit?16:05
CarlFK[m]I am getting better at remembering16:05
CarlFK[m]um.. look in the bar at the bottom16:05
jeaah, found it16:05
CarlFK[m]^ax or xz or something16:05
jeaclosed minicom, ran the command again, still fails :(16:06
CarlFK[m]any chance  hdmi2usbd is running?16:07
jeait shouldn't be16:07
CarlFK[m]Now I forget how to 'fix' it once it is in that wonky state - I thought just closing the term connection was enough, but myabye not16:07
CarlFK[m]sec.. I should have a gst  command to just display the hdmi2usb stream16:08
jeai have completely reinstalled that machine, so hdmi2usbd better not have appeared from nowhere!16:08
CarlFK[m]crap.. can't find it..  or the mplayer command16:13
CarlFK[m]you have something .. cheese? that wil display the v4l stream ?16:14
CarlFK[m]just plugged in my opsis16:18
CarlFK[m]encoder: 1280x720 @ 0fps from pattern (q: 85)16:19
jeayeah, mine says the same thing16:20
jeai think something is wrong with the status16:20
CarlFK[m]well.. 'good news' is mine works16:20
CarlFK[m]mithro.. halp16:21
CarlFK[m]voctomix-ingest --video-source hdmi2usb --video-attribs device=/dev/hdmi2usb/by-num/all0/video16:21
CarlFK[m]I did that, it came up in voctogui16:21
jeado you know what firmware yours is using?16:21
CarlFK[m]         built: Jan 15 2017 19:08:4216:21
CarlFK[m]  version16:22
tpbTitle: Ubuntu Pastebin (at
CarlFK[m]did you flash from the prebuilt repo?16:23
CarlFK[m]mithro (IRC): it is tomorrow.  symlink16:23
jeayes, this is flashed from the prebuilt repo16:30
jeai don't have cheese, so can't use that to test16:30
CarlFKI'll try this   (06:05:58 AM) mithro: jea:
tpbTitle: HDMI2USB-firmware-prebuilt/archive/master/v0.0.3-710-g947321b/opsis/hdmi2usb/lm32 at master · timvideos/HDMI2USB-firmware-prebuilt · GitHub (at
jeayou don't have to change just to match mine16:33
jeacould be something else entirely, or if it is firmware, don't want you with a broken one16:33
CarlFKI have a week to fix it16:33
CarlFKcd HDMI2USB-firmware-prebuilt/archive/master/v0.0.3-710-g947321b/opsis/hdmi2usb/lm3216:37
CarlFKhdmi2usb-modeswitch --flash-gateware flash.bin16:37
CarlFKthat look right?16:37
jeaooh, i managed to see a flash of the opsis when running the command manually from the shell, then it crashed with the broken pipe and the picture went away16:38
jeaum, i think it is something like that16:38
jeaben did the firmware earlier and I wasn't watching16:39
CarlFKmaybe someday we will have docs16:39
jealooking at the history from tonight, that is the thing you want16:40
CarlFKhdmi2usb-mode-switch is already the newest version (0.0.0+git20161124-2).16:41
CarlFK-bash: hdmi2usb-modeswitch: command not found16:41
CarlFKah... dash I bet16:41
CarlFKFileNotFoundError: [Errno 2] No such file or directory: 'openocd'16:42
jeai have a feeling it is crashing due to whatever is making it keep deregistering and registering as a uvc device16:42
CarlFKthats progress...16:42
CarlFKoh hell16:42
CarlFKor good really16:42
CarlFKI am 90% sure that is the service16:42
CarlFKstop-all... should have stopped that16:42
CarlFKcan you give me $ apt-cache policy openocd16:44
CarlFKI get         500 stretch/main amd64 Packages16:44
jeai have the same, but using instead of deb.debian.org16:45
CarlFKcool -I was expecting a ppa16:45
CarlFKand that is the box that flashed?16:45
jeathat is from the machine i just installed this all on16:46
jeanot the one ben did the flashing on - that was an ubuntu machine16:46
CarlFKcan you see where that one came from16:46
jea(which I can't get to because it is currently becoming another mix machine)16:46
CarlFKmithro: will you PLEASE PUT UP SOME DOCS!!16:47
CarlFKi hate that we all just flounder around until maybe it works16:47
jeayes, i agree16:47
jeashould i run stop-all, then start voctocore manually?16:48
CarlFK#[email protected]#!!!16:48
CarlFKor tell me when you do stuff16:48
jeaok. i have stopped everything, then started voctocore manually. it is now in a loop trying to do something16:50
CarlFKwhat is in a loop?16:50
jea(i started with -vvvv)16:51
CarlFKmy guess is something is trying to connect to it and failing.. and the someting is in a loop16:51
jeai wonder what else16:51
CarlFKcheck dmesg, see if the uvc thing is still reg/dereg16:51
jealet me check16:52
CarlFKor.. do you have the grabber box on the lan?16:52
jeayes it is16:52
jeaoh, could be. i'll see if it is on16:52
CarlFKtumbleweed: halp.. stop-all-the-things isn't16:52
jeaok, other things are disconnected from the network16:53
jeastill seeing stuff in dmesg, and voctocore is still doing its loop16:54
CarlFKdmesg stuff is coming from the local .. thing16:54
CarlFKthat I don't know how to control16:54
CarlFKstop-all-the-things is a small script that does something.. have a look, see if you can figure out what really needs to be done16:55
CarlFKIm trying to figure out the openocd thing16:55
CarlFKjea: 3am - how much longer are you up?17:07
CarlFKI have a ubuntu box.  that decided to only show the boot screen then turn it self off17:08
CarlFKopened the case, now it works17:12
CarlFKof course.17:12
jeawell, i am now at the point of deciding whether to use any of this to record stuff today, or just do a camera17:12
jeafinally got opsis to vocto17:14
jeasudo systemctl stop videoteam-ingest17:14
jeathen run the command manually17:14
jeai would love to know why it does not work from the service17:15
jeainteresting it seems to be the only part run as a root user17:15
jeaok, everything just became viable again17:16
CarlFK[m]I know that feeling17:18
CarlFK[m]stop-all-the-things was a last min hack.  not surprised it isn't solid17:19
CarlFK[m]ima try to flash the bios on this box :D17:20
jeai think the ingest service needs to become a user service17:20
jeathen it might work17:20
CarlFK[m]woot.  bios flashed. (now all 4 i5 boxes are 'right'17:27
CarlFK[m]seems I needed new bios to make upgrading i3 to i5 work17:27
CarlFK[m]unless I want a box that just sits and beeps17:27
jeai had to do similar recently17:27
jeadifferent generations of i processors needed a new bios17:27
CarlFK[m]now it post's super fast17:29
jeahmm. i think the ingest issue might be due to the usb audio device17:31
jearunning it manually I get: "Device 'hw:1,0' is busy"17:31
CarlFK[m]that makes some sense17:34
CarlFK[m]how are you getting audio?17:34
jeathe intention is to do like at LCA, with a usb mixer coming in17:35
jeafor lca (and in mine) they seem to use alsa_device to choose which thing17:36
tpbTitle: av-foss-stack/r4mix.yml at master · xfxf/av-foss-stack · GitHub (at
jeai have no idea why the device would be busy though17:37
CarlFK[m]my guess is pulse or something took over17:40
jeashould i be killing pulse?17:42
jeaI remember it coming up in the past, but I don't remember what is done17:42
CarlFKI think so17:43
tumbleweedCarlFK: what did it do?17:43
tumbleweed(or not do?)17:43
* jea waves to tumbleweed 17:44
CarlFKtumbleweed: seems stop--- isn't stoping the opsis service17:44
CarlFKbut hold that..17:44
CarlFKright now audio sound usb.. halp17:44
tumbleweedurgh, I try to stay clear of that17:44
CarlFK(12:31:30 PM) jea: running it manually I get: "Device 'hw:1,0' is busy"17:44
tumbleweedbut we did use it at lca...17:44
tumbleweedis pulse running?17:44
jeai worked out that the stop script doesn't work due to sudo17:45
jeayes, pulse is running. i'll kill that17:45
tumbleweedhrm, we have stuff in the ansible repo that tries to get rid of pulseaudio, IIRC17:45
CarlFKit is a debian desktop install - um, "nothign selected" or sething that I don't understand17:45
tumbleweedright, so you probably want to get rid of some stuff17:45
tumbleweedand/or use our session17:46
jeapossibly. i used the desktop installer, but only selected the standard package when it came up (no DE, etc)17:46
tumbleweedright, we assume minimal installs17:46
jeaso from what I can tell that should be the same as the preseed that you use, but i could be wrong17:47
tumbleweederr desktop installer always installs DE, doesn't it?17:47
jeadidn't seem to17:47
jeabecause I unticked all the DE options17:47
tumbleweedoh, Debian, not Ubuntu17:47
tumbleweedI assume ubuntu in this channel :P17:48
jeanormally you would be correct!17:48
jeabut I thought I would be good and use debian :P17:48
tumbleweedthe graphical installer is just a different frontend to the text-based one17:48
tumbleweedit does the same things17:48
jeaok, that is what i was hoping17:48
jeajust saved me a lot of time downloading stuff17:48
CarlFK[m]jea: how did you use the preseed settings?17:48
jeai read them all, then selected the same thing in the installer17:49
* tumbleweed tries to recall his pulseaudio strategy17:49
CarlFK[m]there are some kernel params, like auto=all or something, that may also be causing things to be different17:49
tumbleweedoh, looks like we used pulse17:50
jeahmm, i don't remember seeing the kernel params in the preseed, but i might have missed them17:50
tumbleweedthey're in the thing that runs the preseed :P17:50
tumbleweedi.e. the bootloader17:50
jeaoh, of course17:50
jeathat is why i didn't see them17:50
tumbleweedjea: what's wrong with the sudo-usage in stop-all-the-things?17:51
jeathe user wasn't in sudoers17:51
tumbleweedit should probably abort if you're already root - it expects to do stuff as non-root17:52
tumbleweedblame carl for that, he pulled sudo mangement out of the ansible repo :P17:52
tumbleweed(I think?)17:52
CarlFK[m]maybe - it in the preseed17:53
tumbleweedI can't find anything to back that up, though17:53
tumbleweedat any rate, we're trusting the installer to set that up17:53
CarlFK[m]is sounds like something I would do17:53
jeai can live with that. i'll investigate after this weekend17:53
tumbleweedOK, anything you need me to help with?17:53
jeai'm still trying to sort out the audio17:54
tumbleweedis hw:1,0 correct for your input?17:54
tumbleweedyou may need to change that17:54
jeaif we used pulse at LCA, and we had the USB audio device in the ingest, then I think it shoudl be alright17:54
jeaI'll double check17:54
jeaaplay -l lists only 1 USB audio thing, as card 1 device 017:55
jeaso I think that would be hw:1,017:55
CarlFK[m]"busy" makes me think some other process is locking it17:56
CarlFK[m]is the ingest process something that can be turned off across reboots?   and then turned back on17:59
jeai can start it manually if needed18:01
jeabut i will need audio from something18:01
jeahmm, i managed to get it to start18:02
CarlFK[m]im thinkign a reboot is in order18:02
tumbleweedCarlFK: sure (re turning it off)18:03
tumbleweedsystemctl uninstall videoteam-ingest18:04
tumbleweederr .path and .service probably18:04
jeawell, i think i have enought stuff worked out now, that i may as well sleep for a bit18:05
jeathanks very much both of you for the help. i really appreciate it18:05
tumbleweedgood luck18:05
CarlFK[m]oh hey.. I only have 3 HD cams for PyOhio...18:14
CarlFK[m]hmm... I did test DV into vocto...18:14
CarlFKone room gets PyCon style18:17
tumbleweedI used to pooh-pooh HD conference video as unnecessary, but now SD just looks so dated :/18:19
CarlFKyou have time to think about updating hdmi2usb-mode-switch  package to master ?18:21
tumbleweedI'd want to test it first, which I can't do at my desk at work18:22
CarlFKdon't suppse you can walk me though testing it.. I have it checked out.. the readme gives me hives18:26
jea[m]Anyone have any idea why voctomix ingest would crash when I pass it audio, but it works without audio?23:28
tumbleweedit's not outputting anything?23:29
tumbleweedand vocto isn't saying anything?23:29
CarlFK[m]jea: oh my,..usb audio device right?23:29
jea[m]Yes USB audio23:29
CarlFK[m]sec, leme plug one in23:30
CarlFK[m]without audio = no --audio-source right ?23:30
jea[m]The ingest command fails with a gstbasesrc.c file line23:30
CarlFK[m]not 0 audio level23:30
jea[m]I had the other machine working this morning, but now the second machine is failing :(23:31
CarlFK[m]I hope it doesn't matter, but plugged into usb2 or 3 port?23:33
tumbleweedit shouldn't23:33
tumbleweedusb2 has loads of bandwidth (for audio)23:33
CarlFK[m]I was thinking driver glitch23:33
CarlFK[m]I am sure the audio device is usb223:34
jea[m]I’m not sure I even have USB 2 ports on this desktop23:34
CarlFK[m]do you have the ingest line handy?23:35
jeaError details: #1: gstbasesrc.c(2950): gst_base_src_loop (): /GstPipeline:pipeline0/GstAlsaSrc:audiosrc:23:35
jeastreaming stopped, reason not-negotiated (-4)23:35
jeathat is the error23:35
jeaingest is:23:35
jeavoctomisx-ingest --video-source hdmi2usb --video-attribs device=/dev/hdmi2usb/by-num/all0/video --audio-source alsa --audio-attribs device=hw:2,023:36
tumbleweedare the server audiocaps correct?23:36
jeaum, what is that?23:36
tumbleweedvoctomix's config23:36
jeai haven't touched that, but it could be wrong23:37
CarlFK[m]we haven't really touched that in ... ever :p23:37
jeawhich voctomix config file should i look at?23:37
tumbleweedvoctocore.ini, which ansible writes23:38
CarlFK[m]etc/[email protected]:~$ cat /etc/voctomix/voctocore.ini23:38
tpbTitle: Ubuntu Pastebin (at
CarlFK[m]better be the same23:38
CarlFK[m]jea: try with just audio: voctomix-ingest --port 10001 --audio-source alsa --audio-attribs device=hw:2,023:39
CarlFK[m]that ran for 10 seconds23:39
CarlFK[m]I ^c ed to kill it23:39
jeathey look the same23:40
tumbleweedand the usb audio hardware is the same?23:41
CarlFK[m]aborted right away?23:41
jeayes, aborted immediately23:41
* CarlFK[m] sent a long message: CarlFK[m]_2017-07-21_23:40:53.txt <>23:41
jeai suspect they will be different devices23:41
CarlFK[m]er... you have 2 systems - same over there ?23:41
jeamy second machine worked last night. but it is a different room, and i can't quickly check23:42
jeaif it is useful, i know that the device is working on the machine, because if I let pulse manage it, you can see the audio coming in23:43
CarlFK[m]same usb device?  (guessing the small white 2 channel mixer, usb out23:43
jeait is not the same as what we have at lca23:44
CarlFK[m]k - Im not sure it matters.  usb audio seems pretty standard23:44
jeathis is a yamaha AG03 mixing console23:44
jeayep, that was my thinking23:44
CarlFK[m]how easy would it be to more the audio device to the grabber machine ?23:46
jeaum, a bit difficult23:47
CarlFK[m]168     elif args.audio_source == 'pulse':23:48
CarlFK[m]169         audio_src = """23:48
CarlFK[m]170                 pulsesrc {attribs} {base_audio_attribs} name=audiosrc !23:48
CarlFK[m]worth a shot, but you will need to figure out the --audio-attribs {attribs}23:48
jeahmm, that would be good to try23:50
jeai will give that a go at the next break23:50
jeawe are capturing everything separately for now. yay23:50
CarlFK[m]somewhere I have the args...23:51
CarlFK[m] --audio-attribs device=alsa_input.usb-Burr-Brown_from_TI_USB_Audio_CODEC-00.analog-stereo23:52
CarlFK[m]--audio-source pulse   --audio-attribs device=alsa_input.usb-Burr-Brown_from_TI_USB_Audio_CODEC-00.analog-stereo23:54
jeaok, i will try that soon23:55
CarlFKpacmd list-sources | grep -e device.string -e 'name:'23:58
tpbTitle: Ubuntu Pastebin (at
CarlFKto get the name of your device23:59
CarlFKhow long till the break ?23:59

Generated by 2.13.1 by Marius Gedminas - find it at!