tconnectsync icon indicating copy to clipboard operation
tconnectsync copied to clipboard

After ~24 hours of running in auto-update mode, stops recognizing new t:connect treatments

Open jwoglom opened this issue 3 years ago • 5 comments

Opening this issue to track an issue I have noticed where after ~24 hours of consecutive runtime, I need to restart tconnectsync in order for it to pick up new t:connect treatments.

The "Last CGM reading from t:connect" value mentioned in the logs stays the same on each cycle, but on restart of the app it immediately picks up several hours worth of data that it previously did not detect.

I will investigate to see why this is happening.

jwoglom avatar May 03 '21 18:05 jwoglom

Just dropping a line that I'm experiencing the same.

Is there any way to detect when it happens?

I'm about to set it up in supervisor, and I'm planning to setup a secondary cron job to restart it every 20 minutes or so.

Thanks!!

JasonJShuler avatar Oct 12 '21 21:10 JasonJShuler

There is some logic present inside auto update which is supposed to terminate the script whenever there haven't been updates for a long time. If you're using supervisor, that would then trigger the script to be re-run, hopefully fixing the problem.

It's entirely likely that this isn't fully good enough, or doesn't trigger in every situation. And it's just a workaround for what is certainly a rare bug that still occurs sometimes. Feel free to try tweaking the AUTOUPDATE_* setting values and see if they help before breaking out cron.

jwoglom avatar Oct 20 '21 15:10 jwoglom

This was very annoying. Similar to @JasonJShuler , I enabled auto restart in docker and use a cron job to stop the running container every 8 hours. I look forward to resolution in the code though.

kakridge avatar Feb 05 '22 19:02 kakridge

I was seeing this when running v0.6.3 via supervisor in a linux VM. It seems to stop syncing around midnight each night regardless of what time of day I restarted it. However, it has sync continuously for over 24 hours since I upgraded to 0.6.5. So the potential fix in 0.6.4 may have done the trick!

kswick avatar Feb 11 '22 02:02 kswick

This was very annoying. Similar to @JasonJShuler , I enabled auto restart in docker and use a cron job to stop the running container every 8 hours. I look forward to resolution in the code though.

I honestly have no idea if it's even working anymore. I sort of stopped caring after my endocrinologist screwed up my prescription and then refused to give me enough insulin I figured if she doesn't care why should I right?

Do keep in mind that this is tandems bug and it's tandem who deserves any kind of frustration. This thing is entirely reverse engineered and @jwoglom is staying on top of it like a champ

JasonJShuler avatar Feb 11 '22 05:02 JasonJShuler