As I said before: I do not run with any limits.
I will do that, when I see the issue the next time.
Not the limiter then, I think. 
I’d probably start with STTRACE=model. You can enable this in the GUI too, when things are slow, without requiring restart. It shows quite a lot about what it’s doing when syncing, but not necessarily exactly what’s taking time. You can redact file names reasonably (i.e., keeping unique things unique) before posting.
Thanks, that is what I was asking for. I already looked into the different debug logging options, but they were not very clear to me.
I guess logs of the master and the slaves are relevant?
Probably, until we know on which side the issue is caused.
I am probably not able to share something today since I had to restart them due to a bug in my script. I will try to find a test run tomorrow.
I say I suspect limiter because we still wrap the connections, and I recall some infinity handling funkyness.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.