RDM responder test "server error" when max personality is 255 #1508
Comments
|
Hi @majcit , I assume this is also you: Sorry, I saw and fixed a bug related to this, but hadn't had a chance to reply yet. My fix is here: However that isn't the resolution to the error you're getting. What version of OLA are you running on what OS and how did you install it? I think this is the fix for the bug you reported (but I'd still expect it should be fixed in any recent OS): |
|
Hi Peter, my colleague setup the OLA server a few monthes ago following instructions: https://www.openlighting.org/rdm-tools/rdm-responder-tests/getting-started/#Installation I should say I have no experience about LINUX , after I struggled command line and found out my LINUX version is 7.11 I will try to update OLA but I wonder is there a simple instruction to update , (for dummies!) ? Regards |
|
That's a version of Linux released in 2016: You can find your Ola version at http://:9090/ Apologies, those docs, like a number of our installation ones, could do with an update. Nowadays you're better following the normal steps to install the latest Raspbian: Then just run: You could in theory update Raspbian, but starting from scratch will be easier. If you've got another SD card, you can do the new stuff on the second card, so you've even got something to fall back on if you have an issue for some reason. |
|
ok, I just installed the latest Raspbian and ola-rdm-test. just as you said, very easy thanks so much. now http://10.1.1.12:9090/ola.html works fine , I can do discovery and send dmx data to my LED fixture successfully. but, http://10.1.1.12:9099/static/rdmtests.html does not work :/ did I miss something? edit: |
|
Excellent, glad to hear it! Did you run the OLAd process manually, or was it just working when you browsed to the address? If you run This should make it start up automatically each time you restart the Pi: |
|
I executed your last commands : 9090 works fine, I suspect All IP numbers are 0, but I can discover and change personality of LED fixture, and turn on LEDs using DMX CONSOLE successfully : 9091 not work at all : |
|
I also tried "rdm_test_server.py" command : after this command opens 9099 page , but does not work, I can not select UNIVERSE or RESPONDER from combobox : |
|
What does What about Does hitting Ctrl+F5 help, what does the browser console error log show? It feels like some of the Javascript hasn't loaded properly. This feels like some of the packaging bugs that @yoe fixed haven't made it into Raspbian yet possibly. |
That's a bit odd. If you do |
|
On Wed, Oct 31, 2018 at 02:04:48PM +0000, Peter Newman wrote:
This feels like some of the packaging bugs that @yoe fixed haven't made it into
Raspbian yet possibly.
That's certainly possible. Fixing things in Debian stable (from which
Raspbian pulls) isn't done automatically.
If you think I should push an update to stable because something is
critically broken there, that can be done. But if there's a known
workaround that we can point people to, that's easier to get through the
red tape ;-)
…--
To the thief who stole my anti-depressants: I hope you're happy
-- seen somewhere on the Internet on a photo of a billboard
|
|
ok, after after after, |
|
there is no error message, 9091 not opens at all before 9091 opens after I hope I understand right the meaning of "error log" as you expected, if not please forgive my poor knowlegde about LINUX ! command prompt log is here : is 127.0.0.1:9099 normal at first line above ?
|
|
The This is indeed due to the bug in Debian. I'll see if I can push an update for that; it's pretty stupid. Meanwhile, you can edit |
I was actually meaning within your web browser you should be able to run diagnostic tools or similar and see a console or error log there, however your log had given enough to go on.
Yes, that's intended, we actually bind to 0.0.0.0, so you can access it from any IPs your machine has (e.g. WiFi and wired), but that doesn't make much sense to report, so instead we display the localhost address. |
That will fix it not starting from the init script @yoe although he's also running it manually which is serving up pages but still not working. However this log line is the key: I think this is another consequence of https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=760414 and Debian policy. Specifically https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=790886 , it looks like in both Stretch and Sid, the symlink is from /usr/share/ola/rdm-server/jquery-ui.min.js, but the HTML has only had the version stripped off, not the custom bit (which presumably hopefully just optimised it/removed stuff not being used in that GUI). I don't know how long a release cycle takes for you, but stretch is a no brainer I'd hope. However I'm hoping to release 0.10.8 imminently (I just need to do the NEWS, although I may add the other spelling of writeable too as I'd promised before), so it might possibly make sense to wait a bit for whichever one a new release would appear in. You should be able to test the test server functionality on a test box by patching the dummy plugin to a universe and then running the test server against one of our dummy responders (they should pass all the tests of course!). |
|
I installed an older version of Raspbian V9 : no succeed :/
for Enttec Usb Pro Device : for Dummy Device : |
Sorry @majcit that was directed at @yoe rather than you.
As was that. If you go back to the latest version of Raspbian and edit |
|
sadly I have to return back and content with OLA 0.9.5 till new release of OLA 0.10.8 I should say binary package image http://dl.openlighting.org/raspbian-ola-0.9.5.zip I kindly request if possible please prepare similar method for new versions. thank you so much |
I tried to run test over Dummy device, I did as following : the result: 9099 opens page,
|
|
FYI, I've just uploaded an update for ola in Debian stable that fixes these issues. It should be part of an upcoming point release. There is one planned for next weekend, but I believe I'm too late for that one already. |
|
👍 İs new update available for users (me) now? |
|
No, you need to wait until it is in Debian stable. You can check the tracking bug in the Debian bts to see when that's done; it will be marked as 'closed' when the updated package is included and available. |
|
(also, if you're using Raspbian, you may need to wait until the update of Raspbian that includes it... Don't know what their exact policies on that subject are, but they're usually reasonably fast with that) |
|
On Wed, Oct 31, 2018 at 07:52:10PM -0700, Peter Newman wrote:
You should be able to test the test server functionality on a test box by
patching the dummy plugin to a universe and then running the test server
against one of our dummy responders (they should pass all the tests of
course!).
Is this something we can do from the command line? If so, might make
sense to add that to an autopkgtest in the Debian package...
…--
To the thief who stole my anti-depressants: I hope you're happy
-- seen somewhere on the Internet on a photo of a billboard
|
See here for discussion on updating those: In short I'm going to prioritise writing code myself, but others are welcome to work on building Pi images and keeping them up to date. |
Can we see the browser's console log for that bit please @majcit ? I take it you had added the tests too, so they move from right to left as per your screenshot?
Just to confirm @yoe did you actually run tests against the dummy responders to confirm the run button (and test results) worked too?
Yes, and yes, was on my list (and still is), maybe doing a basic one that just tests one responder would be a good start (although wouldn't catch this specific issue unless we do webpage stuff too). We need to launch olad (how do we do that in autopkgtest) then run: http://docs.openlighting.org/ola/man/man1/rdm_responder_test.py.1.html |
Today I reinstalled a fresh Raspbian V9 and OLA V0.10.3 from scratch and repeat the instructions again: same result : RUN button does not start. :/ here is the log : |
|
Please can we see the browser's console/error log @majcit ? E.g. developer tools. If you let us know which browser, we can give you some assistance. There will be some sort of Javascript error occurring when you click the button. |
|
I use Chrome , Win7 , as I find instruction from internet , the error log file must be created in as below : BTW I noticed when I changed with original after change to |
You want this one @majcit : Yes, the styling would be intentional and shows the JS is now loading. |
|
ok , finally I learned how to open error log for Chrome. and here is the error log when I press "RUN TEST" button : This error occurs in both if I "add all" test items, also when I dont add any test item. |
|
Excellent, glad you've discovered the joys of the browser console log. @majcit (and indeed @yoe ) these fixes should make it actually work with the later JQuery Debian provides: |
|
thanks Peter , maybe this is silly question but my big question is how other people use new versions of OLA rdm tester, since as I understand >the only< downloadable and working OLA is 0.9.5 |
|
Does that mean I can modify this java script content and change these 11 items to get result? I think I misunderstood the previous post that I have to wait till new release of Raspbian. |
You're only referring to the deb packages from Debian/Raspbian/Ubuntu. Any of those prior to 0.9.1 would have been fine, and possibly anything before 0.10.3.nojsmin-1. Other OSes weren't affected by these bugs (as we bundle our own older JQuery, which works with our code (although could theoretically have issues on newer browsers).
Personally I run the bleeding edge git master version, as that's where I develop and I always like to run the latest tests against any fixture. However as you'll have seen with your initial 255 personalities issue (fixed in eb7ff0d), there is always a chance of an issue slipping through. By it's nature its rather challenging to test a test suite.
Yes, you can even apply this diff file using patch:
You won't need to wait for a full Raspbian release, just for them to update from Debian and release an updated version of that package (once @yoe incorporates these fixes and they make it into the relevant Debian release). Because these files are plain text and interpreted, you can just edit them now and the fixes will work, although if you reinstall it, you'll need to apply the same fixes. |
|
On Sun, Nov 11, 2018 at 08:21:55AM -0800, Peter Newman wrote:
I think I misunderstood the previous post that I have to wait till new
release of Raspbian.
You won't need to wait for a full Raspbian release, just for them to update
from Debian and release an updated version of that package
Well, that does require a full Debian point release, which implies a
full new Raspbian release :-)
(once @yoe
incorporates these fixes and they make it into the relevant Debian release).
Because these files are plain text and interpreted, you can just edit them now
and the fixes will work, although if you reinstall it, you'll need to apply the
same fixes.
Right.
…--
To the thief who stole my anti-depressants: I hope you're happy
-- seen somewhere on the Internet on a photo of a billboard
|
|
thanks for complete explanations. I suceeded updating both : It works fine with dummy devices and my device when Max Personalities < 255 but my device when Max Personalities = 255 "server error" still exists. my device with max personality < 255 passes all tests successfully : same device with max personality = 255 , starts test, at progress bar less than %50 server error occurs : |
Ah, I didn't realise that, I assumed the fixes just appeared when you release the package given the steady stream of updates I get offered on my normal machine (or is this confusion between you releasing a full new version to stable/unstable and a fix getting backported to an older release)?
Excellent, glad you got it sorted.
This isn't the latest working package, it's fixes to make the latest deb package released for your OS function correctly.
Sorry about that @majcit . I probably only tested against our git master and then backported the fix, or possibly only tested the test you'd initially reported as failing. Our git master has had a bit of a rewrite in this area of tests, so it masked that issue (as well as catching if this particular bug happens in our test programming), anyway this will fix that particular (and hopefully final) bug: |
|
ok, and maybe my last question : |
Excellent, glad to hear it!
What does Alternatively you could compile from scratch, or make your own deb or switch to a different OS (e.g. Arch offers git versions: https://aur.archlinux.org/packages/ola-git/ ). So I think I'm right in saying you're happy and this can be closed as far as you are concerned @majcit ? @yoe I assume you'll track merging the outstanding bits to fix oldstable etc in Debian's PTS so you don't need this open either? I'll wait for #1511 to be merged before closing anyway. |
Thanks so much again. I will wait new release 0.10.8 : )
sure I am very happy , I learned new things during this thread. and I got much more than I expected. To setup RDM Responder Tester 0.10.3 on Raspberry Pi :
many thanks to all establishers, engineers and developers of OLA. It is definitly a great tool. |
I encountered an issue during OLA responder test,
my device is RGB LED fixture,
I planned to use all 255 different personalities, (1 DMX mode, 254 DEMO mode)
when I run the OLA responder test with 255 personalities, "server error" occures.
when I reduce personality to 254 (1 DMX, 253 DEMO) test passes sucessfully
according to 10.6.1, personality may be 1-255
so I think this "server error" is related to OLA.
The text was updated successfully, but these errors were encountered: