Make rotctld polling fault tolerant

wose / satnogs-monitor

Terminal UI monitor for your SatNOGS ground station.

GNU General Public License v3.0

78 stars 6 forks source link

Make rotctld polling fault tolerant #29

Open wose opened 3 years ago

wose commented 3 years ago

Hi, I am using SPID RAS rotator and I believe I'm having the same issue. I'd need to slow down the polling of position. I've tried querying position via rotctl manually during an observation and sometimes I even got an error. The response of rotctl is noticeably slow (almost a second) when idle. God knows what's happening in there when the observation is running. I've tried setting up rust, but I could not get it running. I'll give it a shot some other day.

Originally posted by @AdamKrovina in https://github.com/wose/satnogs-monitor/issues/26#issuecomment-833023441

wose commented 3 years ago

Increasing the polling interval didn't help in this case. We should:

[x] check rotcltd code to see what's causing the polling error
[ ] retry getting the rotator position when it failed with an error in the next polling cycle

wose commented 3 years ago

The hamlib part seems ok to me. They forward any error during communication with the rotator as IO Error and swallow errno, which could hold more information on what's happening. The documentation of the rotator implementation states polling the position during movement is ok and it's also synchronized between all clients.

@AdamKrovina: You could try to set additional parameters for hamlib in satnogs-setup -> Advanced -> Hamlib -> HAMLIB_UTILS_ROT_OPTS. Something like -vvvvv should give you more logs, and also print errno. Not sure we can do anything about it though.