Jan Paulsen

Would adding pools instead of hostmasks to a job increase its performance? For example, Berlin: 100, rest: 0? Currently, we have a pool per site, but we don't assign them...

Thanks @timurhai , I was able to add the capacity and it works as intended. 🌹 I still would be suuper happy if you could add the ticket information as...

Ahh, I see the problem @timurhai ... If you find the time for a future release, it would be super helpful for us. Thanks a lot Jan

Hey guys, I am writing to update you guys (@timurhai @lithorus ) with some statistics of how many of these "frozen tasks" we have. @lithorus We made a little cronjob...

@lithorus Uff... i like the idea of the "keep-alive" check you mentioned. This could be worth to try out and to not waste the 30mins/taskNoProgressTimeout of rendertime but to crawl...

@lithorus And here a little very brief guide on how our script works atm: ``` Check every 30 sec each renderhost: if render running: renderTask = fromRender if renderTask.progress ==...

> What is server log this time? What server "thinks" about this render and its tasks? It does not became "zombie" and still runs latest task? And task progress timeout...

Here are our settings: We have 1000 renderclients connecting, ~200 monitors (heartbeat 2,5sec) , ~NIMBY-Trays-Setters (heartbeat 5sec), ~5 get-coronjobs (heartbeat 5sec) cgru/3.3.0/config_default.json ``` "include":["afanasy/config_default.json","config.json"], ``` cgru/3.3.0/config.json ``` { "cgru_config": {...

I will try to set the heartbeat to 2-3 on our different sites and see if it changes the amount of stuck renders my script catches. nothing was set, so...

I suppose an immediate renderservice restart would be the best solution if we cant resolve the problem at its root. Please keep in mind that sometimes other tasks can start...