a nill instance was returned

[dcorsus]

Here is a printout of the full settings page...

WOW ....
How many players?
I already see duplicated players in the table. I guess this was not after you deleted the sonos.ini file, right? How long ago was that? I wonder if this a load/timing thing

[bsobel]

Fyi, looking at the log, I see the GetZoneGroupState 500 errors, if I go to those URL's directly with a browser they render the XML just fine... Not sure what that suggests yet, but FYI...

[bsobel]

WOW ....
How many players?
I already see duplicated players in the table. I guess this was not after you deleted the sonos.ini file, right? How long ago was that? I wonder if this a load/timing thing

I reset the configuration a few days ago, so this is the current sonos.ini. I do see some name errors that come up with stereo pairs (I started a thread about that, don't think you responded to that one yet) No idea why the zones are 'duplicating'...

On the zone count, yeah this house is pushing the limits with Sonos (my others all have < 10 zones). Have been in discussions in the past with their CEO on moving the zone limit to 64, but no luck yet

[bsobel]

On the menu you can see that the duplicates are showing in some but not all cases I think, also (not sure related) the stereo pairs show up (the _RF units). I would think those should collapse into one unit hiding the pair, like the Sonos app, no?

[bsobel]

Here is a thread from the python imp called SoCo (used to use it prior to your plugin), seems to be same issue:

https://github.com/SoCo/SoCo/issues/518

Key takeaway is once the errors occurs, may have to go to the device directly with a get request to get the page instead of relying on UPnP. Maybe a Sonos UPnP bug... Any chance you could consider testing that option, if you receive a 501 fall back to a direct request with the player instead of via UPnP?

[dcorsus]

Here is a thread from the python imp called SoCo (used to use it prior to your plugin), seems to be same issue:

https://github.com/SoCo/SoCo/issues/518

Key takeaway is once the errors occurs, may have to go to the device directly with a get request to get the page instead of relying on UPnP. Maybe a Sonos UPnP bug... Any chance you could consider testing that option, if you receive a 501 fall back to a direct request with the player instead of via UPnP?

I suspect that the 501 error is a Sonos UPNP bug. Wouldn't be surprised that their internal buffer isn't big enough (or the UPNP buffer size is exceeded) because it returns quite a bit of info ....

Unfortunately, I use that info for alarms and pairing info, so that might be an issue.

I'm also wondering whether the duplicate players has perhaps something to do with HS. I see you have about 560 devices belonging to this PI, wonder if there is some other bug lurking somewhere. Long time ago, another user found and issue when # devices being >10000. Another source could be the cashing of .ini by HS, I wonder if it starts failing functions if some of the sections become too large. Need to think how I can figure this out.
I may need a longer term log.

Here's what I might need:
a/ set the pi debug log (not superdebug)
b/ set the log to disk flag
c/ rather than deleting the .ini file use the "delete all players button at the bottom of the config page". This will cause the PI to restart
d/ let it run for say an hour (maybe we'll need to let it run until we see a duplicate player)
e/ turn log to disk off
f/ post log and ini file

Dirk

[bsobel]

Here you go. Saw that the Massage Room device was already duplicated, also noticed that the player table didn't have names for many and IPs for some devices...

[dcorsus]

Here you go. Saw that the Massage Room device was already duplicated, also noticed that the player table didn't have names for many and IPs for some devices...

hmm, when the PI comes on, there are still players in the .ini file, did you do the delete all?
Could you post the .bak log file (which is post to restart).

[bsobel]

Yes, I did the delete all. The plugin exited as you said it would. HS tried to restart it automatically it but I got the 'instance cant be found....' so I restarted HS at that point

Sonos.ini.bak attached.

[dcorsus]

Yes, I did the delete all. The plugin exited as you said it would. HS tried to restart it automatically it but I got the 'instance cant be found....' so I restarted HS at that point

Sonos.ini.bak attached.

hmmmm ..... my worst fears ..... delete-all-players is broken again, something with HS!!!
I brought all my Sonos devices on-line (19) and see immediately issues with delete-all.
I have been doing some moving this afternoon and my dev environment is down. I see that you have duplicated devices now, even more so due to the issues with delete-all.
Suggest you stop hs3, delete the .ini file which is a cleaner way to start fresh (unless we have issues here with HS3 as well).
I did notice that the initialization took for ever and I had some players without names as well, I'll play with my 19 players some more.
I'll post some updates as soon as I can, but I may have to consult Rich who a bit ago made some changes with PI init (and I guess shutdown) and I suspect something got broken.
Sorry for the bad news, stay tuned,
Dirk

[bsobel]

hmmmm ..... my worst fears ..... delete-all-players is broken again, something with HS!!!
I brought all my Sonos devices on-line (19) and see immediately issues with delete-all.
I have been doing some moving this afternoon and my dev environment is down. I see that you have duplicated devices now, even more so due to the issues with delete-all.
Suggest you stop hs3, delete the .ini file which is a cleaner way to start fresh (unless we have issues here with HS3 as well).
I did notice that the initialization took for ever and I had some players without names as well, I'll play with my 19 players some more.
I'll post some updates as soon as I can, but I may have to consult Rich who a bit ago made some changes with PI init (and I guess shutdown) and I suspect something got broken.
Sorry for the bad news, stay tuned,
Dirk

No worries, rather deal with it and find root cause, makes the whole community better. If I nuke everything and restart, ok to cut/paste my grouping information from the INI file so I don't have to recreate those by hand?

[dcorsus]

No worries, rather deal with it and find root cause, makes the whole community better. If I nuke everything and restart, ok to cut/paste my grouping information from the INI file so I don't have to recreate those by hand?

Yes you can cut/paste all linkgroupinfo.
Have email into rjh
Even deleting a single player causes the PI to be terminated by HS before or while entries in the .ini file are being deleted, causing all kind of mayhem.
Tried to turn option to "load PI in background" off but made no difference.

Dirk

[dcorsus]

No worries, rather deal with it and find root cause, makes the whole community better. If I nuke everything and restart, ok to cut/paste my grouping information from the INI file so I don't have to recreate those by hand?

OK, embarrassing but issue of delete single and delete all was in my PI. Thanks to Rich who quickly confirmed everything looked kosher on the HS3 side of the house which motivated me to look a bit harder. So this is going to be fixed soon.
During my debugging I might have seen something else that needs further investigating. I stumbled onto some "instances" array that has a max capacity of 16, so either it adjusts its own capacity (need to test tonight) or having more than 16 players is causing mayhem.

Stay tuned.

[dcorsus]

No worries, rather deal with it and find root cause, makes the whole community better. If I nuke everything and restart, ok to cut/paste my grouping information from the INI file so I don't have to recreate those by hand?

Just a quick update. I did find the issue w deleting a player or all players causing the DB to get out of sync. I sporadically still see an error when the PI terminates but discussing it with Rich from HS, it appears to be a race condition and benign.

I've been doing some more testing and I'm still trying to do a test on pushing the PI to have more than >16 (actually 17) instances to see if there is another issue, but I'm maxed out at 17 players so bsobel, I may need you to make me one more start-up log; I did add some debug statements. I can either post a beta version here or submit through the updater.

Thoughts?

Dirk

[bsobel]

Just a quick update. I did find the issue w deleting a player or all players causing the DB to get out of sync. I sporadically still see an error when the PI terminates but discussing it with Rich from HS, it appears to be a race condition and benign.

I've been doing some more testing and I'm still trying to do a test on pushing the PI to have more than >16 (actually 17) instances to see if there is another issue, but I'm maxed out at 17 players so bsobel, I may need you to make me one more start-up log; I did add some debug statements. I can either post a beta version here or submit through the updater.

Thoughts?

Dirk

Happy to do that and which ever version is easier for you to publish is fine with me...