johnnyt Posted May 27 Posted May 27 I have 3 instances of this plug-in running and was able to install new version for all three instances. The new info did not show up so I did the following, in order: 1. wait a bit. didn't do anything 2. restart each plug-in. nothing 3. re-launched IoX GUI - no joy 4. rebooted hardware - 2 of the three plug-ins now show the new data in IoX What is/are the step(s) needed after a re-install to get new info activated/passed on to IoX? If that fails, what is the step to force updates to be recognized and passed on to IoX? Am on PG3x 3.2.22 and IoX 5.8.3. Thanks
Jimbo.Automates Posted May 27 Posted May 27 None of those steps should be necessary, all you should need is: Update plugin Wait till installation is complete Make sure it shows 1.1.1 in the PG3 UI Start Admin Console, the AC loads the profile from the IoX on startup so it must not be running to see the new info. PG3 handles loading the profile into the IoX so if it's not happening will need help from @bmercier to review.
johnnyt Posted May 27 Author Posted May 27 I restarted the IoX GUI - still nothing for one of the plug-in instances I don't see any data coming for the new nodes (or whatever you call those), e.g. GV6, ST, VOCLVL, with log at "Info" level
Jimbo.Automates Posted May 27 Posted May 27 Please download log package and PM it to me and reference this post.Sent from my Pixel 8 Pro using Tapatalk
Jimbo.Automates Posted May 27 Posted May 27 Driver ST is set by PG3 on startup so you won't see it in the Plugin log. GV6 is only set on startup based on settings in the plugin, so it won't show in the log either. It will show if you can change the Poll Sensor in the AC. In the log you sent I see VOCLVL being set 2024-05-27 11:14:59,999 MQTT udi_interface.interface INFO interface:_message: Successfully set s_2930071732 :: VOCLVL to 1 UOM 96 I reviewed the log and it is pushing the profile to IoX so not sure what the issue is. 2024-05-27 09:16:03,410 Thread-11 udi_interface INFO Controller:handler_start: Started Airthings-Consumer NodeServer 1.1.1 2024-05-27 09:16:03,436 Thread-11 udi_interface.interface INFO interface:updateProfile: Sending Install Profile command to Polyglot.
johnnyt Posted May 27 Author Posted May 27 For what it's worth I tried flushing Java cache and restarting GUI (Admin Console) - still nothing for the one instance. So then I tried a another reboot (eISY) - still nothing. I see in the event viewer the following entries related to the plug-in in question. It seems like an incomplete update. I have 7 devices and just looking at "CO2LVL" and "BARPRES" there are only 4 of each one of those. A little later on there were 6. How come it's not 7 every time? I do see some of the new data values being received by IoX, e.g. GV5 (Seconds Since Seen), so it's an issue with GUI not adding the new element. Is this something I need to report to UDI support? Mon 05/27/2024 02:17:02 PM : [n008_s_2930037297] GV5 76 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930037297] CO2LVL 568 (uom=54 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930037297] BARPRES 9913 (uom=56 prec=1) Mon 05/27/2024 02:17:02 PM : [n008_s_2930037297] GV4 161 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930071732] GV2 1716833656 (uom=151 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930071732] GV5 169 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930071732] CO2LVL 528 (uom=54 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930071732] BARPRES 9913 (uom=56 prec=1) Mon 05/27/2024 02:17:02 PM : [n008_s_2930071732] GV4 106 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930029938] GV5 275 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930029938] GV3 -47 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930073475] GV5 266 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930073475] GV3 -60 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930033257] GV5 287 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930033257] GV3 -72 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930027997] GV2 1716833681 (uom=151 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930027997] GV5 144 (uom=56 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930027997] CO2LVL 526 (uom=54 prec=0) Mon 05/27/2024 02:17:02 PM : [n008_s_2930027997] BARPRES 9915 (uom=56 prec=1) Mon 05/27/2024 02:17:03 PM : [n008_s_2930027997] GV3 -60 (uom=56 prec=0) Mon 05/27/2024 02:17:03 PM : [n008_s_2930027997] CLITEMP 245 (uom=4 prec=1) Mon 05/27/2024 02:17:03 PM : [n008_s_2930027997] GV4 108 (uom=56 prec=0) Mon 05/27/2024 02:17:03 PM : [n008_s_2930206779] GV2 1716833605 (uom=151 prec=0) Mon 05/27/2024 02:17:03 PM : [n008_s_2930206779] GV5 220 (uom=56 prec=0) Mon 05/27/2024 02:17:03 PM : [n008_s_2930206779] CO2LVL 452 (uom=54 prec=0) Mon 05/27/2024 02:17:03 PM : [n008_s_2930206779] BARPRES 9919 (uom=56 prec=1) Mon 05/27/2024 02:17:03 PM : [n008_s_2930206779] GV3 -56 (uom=56 prec=0) Mon 05/27/2024 02:17:03 PM : [n008_s_2930206779] GV4 198 (uom=56 prec=0)
johnnyt Posted May 28 Author Posted May 28 This morning I tried re-installing the plug-in in the slot that isn't showing up in the AC . Didn't fix the issue. Even after restarting both the plug-in and the AC and waiting a bit. Still nothing What is the next step in troubleshooting this?
johnnyt Posted May 28 Author Posted May 28 I noticed re-installing this plug-in overwrites the config. Got message about short poll. Change the poll values then nothing. it stopped updating. Had to restart plug-in to get polling working again. Maybe the restart is needed always after re-install. If so, could that be added to the documentation next time you're updating that? And could any existing configuration not get overwritten? Thanks for considering these things.
Jimbo.Automates Posted May 28 Posted May 28 Re-installing in the same slot should not overwrite the config, I've never seen this happen. Restart is not necessary after a re-install since it stops the plugin, installs it, then starts it up. I've reinstalled on mine with each update an none of that is necessary. But in any case, those would be PG3 issues @bmercier would have to look into it.
johnnyt Posted May 31 Author Posted May 31 I've reported to @bmercier, who asked me to open a support ticket so I did and sent all info I have On 5/28/2024 at 4:43 PM, Jimbo.Automates said: Re-installing in the same slot should not overwrite the config, I've never seen this happen. I think I know why I thought the config had been overwritten. I got a message that my short poll was too low, which I then changed but had been configured adequately all the time. It's because the error message over counts the devices being polled by an instance. I added a device, did a "Discover" and, even though the total number of airthings wave devices is now 8 (9 nodes when you count controller), the message said I had 15 polled sensors when I first did it, then said I had 16 polled devices when I did it again (just to capture warning message for you). See screenshot below. Maybe, partly, it counts all devices controlled by all instances instead of just the ones for which the short poll applies? However that still doesn't quite explain things because I have 13 devices, not 15 or 16. Perhaps it counts the controller as a polled sensor? but why 15 the first time then 16? The good news is that the new device did show up in the Admin Console though not without a restart of of the AC. After this I still, though, don't see in the AC any of the new data fields that were added in v 1.1.1 of the plug-in - including no new data for the new device that was just added.
johnnyt Posted May 31 Author Posted May 31 ok, it's not counting all devices for all instances. i just did a "discover" on instance with 3 sensors, and it reports 7 (not 13+)
johnnyt Posted May 31 Author Posted May 31 maybe this is fixed in v 1.1.2 but today one of my instances reports an extra polling sensor. wasn't like this yesterday. don't think I did anything to make it happen. will try new version shortly
johnnyt Posted May 31 Author Posted May 31 On 5/28/2024 at 4:43 PM, Jimbo.Automates said: Re-installing in the same slot should not overwrite the config, I've never seen this happen. ok, so a reinstall does reset config. here's a screenshot before re-install and after. short poll went from 320 to 60 and long poll from 360 to 240
Jimbo.Automates Posted May 31 Posted May 31 20 minutes ago, johnnyt said: maybe this is fixed in v 1.1.2 but today one of my instances reports an extra polling sensor. wasn't like this yesterday. don't think I did anything to make it happen. will try new version shortly That's exactly what was fixed in the new version if you read the release notes...
Jimbo.Automates Posted May 31 Posted May 31 15 minutes ago, johnnyt said: ok, so a reinstall does reset config. here's a screenshot before re-install and after. short poll went from 320 to 60 and long poll from 360 to 240 As I mentioned, nothing I can do about that, it's on the PG3 side.
johnnyt Posted May 31 Author Posted May 31 33 minutes ago, Jimbo.Automates said: That's exactly what was fixed in the new version if you read the release notes... sorry could not infer that from release notes. thought it could be different than the 16 devices when there are only 7. Very different numbers and also a totally different manifestation. Did not see the second problem shown at all when I reported the first one
johnnyt Posted May 31 Author Posted May 31 After upgrading to 1.1.2 the instance that was not working - and still doesn't work - after the first upgrade, it no longer worked at all. I changed the config, which got reset by the reinstall, and, after no polling was done for > 20 mins, I did a restart of the plug-in. Got the message that it will try to rediscover on next poll. Waited 20 mins, did another restart - same message and no data updates for any of the sensors. I understand that you view this as a PG3x problem and I have a ticket open about it with UDI but it is happening to your plug-in so I thought you should know. I also am pointing to this forum thread so this will hopefully inform @bmercierand/or anyone else looking into it. Did a "Discover" and things came back (though still no new 1.1.x data fields in AC). While this is not a major problem, things are not restarting properly by themselves and, given they're supposed to, I thought this would be important to know.
Jimbo.Automates Posted May 31 Posted May 31 15 minutes ago, johnnyt said: I understand that you view this as a PG3x problem and I have a ticket open about it with UDI but it is happening to your plug-in so I thought you should know. I also am pointing to this forum thread so this will hopefully inform @bmercierand/or anyone else looking into it. It helps to tag the person, otherwise I assume you are talking to me. I've never seen this issue before, and it's out of my control so doesn't help to keep on it. Post to the ticket. But, I would suggest moving it to a different slot, something is not right with that one, or delete all the nodes in the PG3 UI Nodes page and restart the node server, but that might not help either. 16 minutes ago, johnnyt said: Did a "Discover" and things came back (though still no new 1.1.x data fields in AC). While this is not a major problem, things are not restarting properly by themselves and, given they're supposed to, I thought this would be important to know. The "will try to discover on next poll" is because you really pissed off the Airthings servers and they are not allowing access. I did this earlier today because I had my production and test servers running at the same time and it took almost 40 minutes before it gave me access again.
johnnyt Posted June 1 Author Posted June 1 1 hour ago, Jimbo.Automates said: The "will try to discover on next poll" is because you really pissed off the Airthings servers and they are not allowing access. It might have helped if the re-install hadn't changed the short poll from 320 down to 60 and, better yet, kept the polling cycle the same, which it does not show any sign of doing (and might be hard thing to do, I presume). Most restarts of my 8 device controller instance will cause api rate limit to be exceeded, especially since it had to be restarted twice - once for the reinstall and again after I returned config to what it was supposed to be to avoid the api rate limit. It's a bit strange that the other controllers reported that too. They only have 2 and 3 devices, respectively. But then again I had to change their config back too. So just came too quickly, I guess. I have to say I've never seen this message before now.
johnnyt Posted June 4 Author Posted June 4 I learned the offending instance of the plug-in has 2 sets of profile files, one in lowercase and one in uppercase. It looks like a previous version of the plugin was using uppercase, and the latest version - with new data fields - is using lowercase. IoX will use the uppercase first, then lowercase if uppercase not found. http://192.168.100.249:8080/rest/profiles/ns/8/files/ <profiles> <profile family="10" id="8"> <files dir="nls"> <file name="EN_US.TXT"/> <file name="en_us.txt"/> </files> <files dir="nodedef"> <file name="nodedefs.xml"/> <file name="NODEDEFS.XML"/> </files> <files dir="editor"> <file name="editors.xml"/> <file name="EDITORS.XML"/> </files> </profile> </profiles> I thought it strange that I was only having issues with one instance but this was the initial instance I installed and was the only one running for a while - so probably an issue that's been there since the first version of the plug-in. The only fix I'm been offered is to delete the plug-in and install it fresh. This worries me as I do have a lot of dependent programs. I went to try deleting an instance that has the fewest programs to see if/how deleting then installing (back in same slot) will restore all programs. I get this worrisome message that could mean all three instances will be gone if I click "Confirm". @Jimbo.Automates, have you ever tested deleting just one instance of the plug-in when you had a couple more running? Does it uninstall all instances or just the one for which you clicked the Delete button? Does installing back in the same slot restore all the program dependencies? Any info would be appreciated.
Jimbo.Automates Posted June 4 Posted June 4 21 minutes ago, johnnyt said: I learned the offending instance of the plug-in has 2 sets of profile files, one in lowercase and one in uppercase. It looks like a previous version of the plugin was using uppercase, and the latest version - with new data fields - is using lowercase. IoX will use the uppercase first, then lowercase if uppercase not found. As I discussed with Benoit, these have never changed, I've had plugin installed from day one on my production system, with no issues. We have no clue how this could have happened. You can see the full history of the editors file and it was never uppercase so something funky happened at some point on the PG3 side. https://github.com/UniversalDevicesInc-PG3/udi-poly-Airthings-Consumer/commits/main/profile/editor/editors.xml This was exactly why I told you a while back to delete or move to another slot. 22 minutes ago, johnnyt said: The only fix I'm been offered is to delete the plug-in and install it fresh. This worries me as I do have a lot of dependent programs. I went to try deleting an instance that has the fewest programs to see if/how deleting then installing (back in same slot) will restore all programs. I get this worrisome message that could mean all three instances will be gone if I click "Confirm". @Jimbo.Automates, have you ever tested deleting just one instance of the plug-in when you had a couple more running? Does it uninstall all instances or just the one for which you clicked the Delete button? Does installing back in the same slot restore all the program dependencies? Any info would be appreciated. Deleting plugins is part of PG3 that I don't own, but pretty sure it will only delete the one in the slot you request to be deleted. But, as always, if you are not sure, make a backup of IoX and PG3. As for programs still working, yes they will still function. Delete the plugin from that slot, reinstall it, then reboot IoX and all programs should be fine.
Recommended Posts