HBlink crash (memory leak)?

[K2IE]

2nd time in one week.

ERROR 2020-11-02 09:49:48,212 (GLOBAL) STOPPING REACTOR TO AVOID MEMORY LEAK: Unhandled error in timed loop.
[Failure instance: Traceback: <class 'KeyError'>: 'LAST'
/usr/local/lib/python3.7/dist-packages/twisted/internet/base.py:1283:run
/usr/local/lib/python3.7/dist-packages/twisted/internet/base.py:1292:mainLoop
/usr/local/lib/python3.7/dist-packages/twisted/internet/base.py:913:runUntilCurrent
/usr/local/lib/python3.7/dist-packages/twisted/internet/task.py:239:call
--- ---
/usr/local/lib/python3.7/dist-packages/twisted/internet/defer.py:151:maybeDeferred
/opt/hblink3/bridge.py:203:stream_trimmer_loop
]
INFO 2020-11-02 09:49:48,219 (REPORT) HBlink reporting client disconnected: IPv4Address(type='TCP', host='127.0.0.1', port=51548)

[K2IE]

Another crash this afternoon. I have rolled back hblink3 to commit 08f6669 which is where I had it for some months and it was completely stable. Will monitor to see if this makes any difference.

[picchiosat]

Hi everyone, the Memory LEAK problem is already present since private calls were implemented. I had already reported to Cort about the problem (see https://dvswitch.groups.io/g/HBlink/message/1735). Blocks appear to come during incomplete transmissions of SMS or GPS locations. (usually to a private call)

[randybuildsthings]

I will keep this in the mix for now, as it needs to be looked at. Generally speaking, having the reactor kill itself rather than exhaust all memory in the system is admittedly a better solution for the moment, but detecting that a memory leak condition is happening is not better than avoiding it or correcting it as you go. So, I'll keep this open until there's time to look at the nature of the memory leak and if it's possible to set things up so that resources can be cleaned up and deallocated, so as to avoid the memory leak condition.