I'm not sure if this was considered in the design or is expected.
Anyway, when my host restarted unexpectedly today, more than half of the selkies containers had problems, which were manifested in repeated retries of wss in the WebUI or failure to obtain http resources.
Finally, after I deleted all non-app data in the config dir, the problem was solved.
Although not sure of the specific issue, it looks like some status files in the config dir are corrupted.
I think containers would be better off keeping as little dependency on session state as possible, or needing better recovery mechanisms.
Anyway, this is my feedback on what I encountered today.
I'm not sure if this was considered in the design or is expected.
Anyway, when my host restarted unexpectedly today, more than half of the selkies containers had problems, which were manifested in repeated retries of wss in the WebUI or failure to obtain http resources.
Finally, after I deleted all non-app data in the config dir, the problem was solved.
Although not sure of the specific issue, it looks like some status files in the config dir are corrupted.
I think containers would be better off keeping as little dependency on session state as possible, or needing better recovery mechanisms.
Anyway, this is my feedback on what I encountered today.