Maintenance
Maintenance
I have uploaded an updated version of the Documentation to the file server at https://wmtprojectsforum.altervista.org ... ls_Players.
Terry
Re: Maintenance
Terry
I had a text from Nick just before lunch to say that the Minster was quiet today, I went back to ask whether the bells / quarterback was ok, sadly he hadn't been near it to be able to confirm or otherwise, I'm assuming he didn't use the web interface to try to kick things off.
I'm in tomorrow so will be able to report anew then, in the meantime can you log in to take a look.
Penri
I had a text from Nick just before lunch to say that the Minster was quiet today, I went back to ask whether the bells / quarterback was ok, sadly he hadn't been near it to be able to confirm or otherwise, I'm assuming he didn't use the web interface to try to kick things off.
I'm in tomorrow so will be able to report anew then, in the meantime can you log in to take a look.
Penri
Re: Maintenance
Penri,
I just did a minor modification to the code which might do the trick. After rebooting, I spoke to Mandy who confirmed that the music was playing.
If this minor fix doesn't work I'm going to have to make another change so that the program always starts with Playlist 1.
I just did a minor modification to the code which might do the trick. After rebooting, I spoke to Mandy who confirmed that the music was playing.
If this minor fix doesn't work I'm going to have to make another change so that the program always starts with Playlist 1.
Terry
Re: Maintenance
Terry
I popped in on my way to town and can report that the music wass playing and the Quarterjack rang his bells.
Thank you
Penri
I popped in on my way to town and can report that the music wass playing and the Quarterjack rang his bells.
Thank you
Penri
Re: Maintenance
Yesterday I added some logging code to the Music Pi software to try to diagnose the problem. The latest version of the code is at https://gitlab.com/wmtprojectsteam/mins ... ngineering.
I monitored the Minster-Music.log all day yesterday and things seemed to proceed as expected and then at just after 5 pm I lost the connection to the Pi. The VPN was OK because I was still able to log in to River System Pis.
This morning, in conjunction with Penri, I monitored the system again and found that prior to 10 am I had no access to the Pi and the Web App didn't work for Penri. At 10:10, there was still no access from VPN or the Web App so Penri power cycled the system. This didn't start the Music, but it did allow network access. The logs for yesterday and today are attached and from these it can be seen that the Pi continued to work overnight. The 6 minute apscheduler triggers can be seen all night and entry to the MP3 startup function can be seen by the line containing '10 am Startup Triggered'. However the music did not start.
There is also some rather strange behaviour in the log for this morning in that the 6 minute apscheduler triggers are reported every hour until 10:06 and then logging begins again at 09:17!
To me it is clear that last night at least network access was lost from the moment that the system shut down for the night. However, that hasn't happened in the past so this may be another problem. To me the key to the original problem is that the music isn't starting at 10:06, even though the scheduledmp3() has clearly been entered. I think that the problem is that the mp3_player_stop() function sets the playlist_index variable to 0 and it isn't therefore passing the if statement in line 258. That's easily fixed, but the network issue is a bit worrying.
Today I'll add the code to restore the last Playlist to the scheduledmp3() function which should solve the original problem. I'm also going to add some more logging code, both in this program and also in the Bells Pi code to try to establish why network access was lost.
I monitored the Minster-Music.log all day yesterday and things seemed to proceed as expected and then at just after 5 pm I lost the connection to the Pi. The VPN was OK because I was still able to log in to River System Pis.
This morning, in conjunction with Penri, I monitored the system again and found that prior to 10 am I had no access to the Pi and the Web App didn't work for Penri. At 10:10, there was still no access from VPN or the Web App so Penri power cycled the system. This didn't start the Music, but it did allow network access. The logs for yesterday and today are attached and from these it can be seen that the Pi continued to work overnight. The 6 minute apscheduler triggers can be seen all night and entry to the MP3 startup function can be seen by the line containing '10 am Startup Triggered'. However the music did not start.
There is also some rather strange behaviour in the log for this morning in that the 6 minute apscheduler triggers are reported every hour until 10:06 and then logging begins again at 09:17!
To me it is clear that last night at least network access was lost from the moment that the system shut down for the night. However, that hasn't happened in the past so this may be another problem. To me the key to the original problem is that the music isn't starting at 10:06, even though the scheduledmp3() has clearly been entered. I think that the problem is that the mp3_player_stop() function sets the playlist_index variable to 0 and it isn't therefore passing the if statement in line 258. That's easily fixed, but the network issue is a bit worrying.
Today I'll add the code to restore the last Playlist to the scheduledmp3() function which should solve the original problem. I'm also going to add some more logging code, both in this program and also in the Bells Pi code to try to establish why network access was lost.
- Attachments
-
- Minster-Music_2021-04-14.log
- (13.03 KiB) Downloaded 53 times
-
- Minster-Music.log
- (11.56 KiB) Downloaded 54 times
Terry
Re: Maintenance
According to the logs. Everything is fine this morning. Mandy has confirmed that the music is playing.
Terry
Re: Maintenance
Earlier this morning I logged into both Pis and checked the logs. I noticed that the Bells Pi was logging a complete shutdown every hour and identified what I believe was the problem. I applied the fix and rebooted the Bells Pi to start the new version of the code. The problem was that I was then locked out of the Bells Pi. 
About half an hour ago I rang WMT and Mandy power cycled the system for me. This appeared to fix the problem and the system appears to be working correctly now.
However, this is the third time that network connectivity has failed; each time under different circumstances. For the most part a failure isn't necessarily a total disaster. The first time it happened was at around about 5 pm when the system is shut down. The Music and Bells Pis continued operating normally part from there being no access to the system from the Web App. The second time was similar although it occurred during normal Opening Times. The third time (this morning) was a bit different in that access to the Music Pi remained good, but I couldn't access the Bells Pi. The Music Pi log indicated that the system was unable to establish the Bells Pi socket connection so the Wedding Bells wouldn't have been able to work (there is a message passed from the Music Pi to the Bells Pi when the Wedding Sequence is almost complete). The Web App would have worked apart from control or monitoring of the Bells Pi.
This is getting a bit worrying. I will keep an eye on this, but I'd like to know what is causing the issue. I will post a query to the Linux User Group.
About half an hour ago I rang WMT and Mandy power cycled the system for me. This appeared to fix the problem and the system appears to be working correctly now.
However, this is the third time that network connectivity has failed; each time under different circumstances. For the most part a failure isn't necessarily a total disaster. The first time it happened was at around about 5 pm when the system is shut down. The Music and Bells Pis continued operating normally part from there being no access to the system from the Web App. The second time was similar although it occurred during normal Opening Times. The third time (this morning) was a bit different in that access to the Music Pi remained good, but I couldn't access the Bells Pi. The Music Pi log indicated that the system was unable to establish the Bells Pi socket connection so the Wedding Bells wouldn't have been able to work (there is a message passed from the Music Pi to the Bells Pi when the Wedding Sequence is almost complete). The Web App would have worked apart from control or monitoring of the Bells Pi.
This is getting a bit worrying. I will keep an eye on this, but I'd like to know what is causing the issue. I will post a query to the Linux User Group.
Terry
Re: Maintenance
Sounds a bit concerning. I have replied to your LUG post with some suggestions. I can go in physically and make the changes if you want, but maybe in a week or so cos I need to adjust to being back home again (was at the hospital for a bit).
Hamish