Open Vallefor opened 4 years ago
is it 1 or is 5 (minutes) ?
@cdujeu I thought that cycle have a random seconds. But then I noted that it is exactly 65 seconds.
I change my nginx config to:
...
location / {
grpc_pass grpcs://192.168.1.49:8013;
client_body_timeout 600s;
}
...
http {
...
keepalive_timeout 600s;
...
}
And cycle become 10 minutes.
If I increase timeouts more than 600s it have no effect.
ha ha interesting the grpc connection does have an inherent 10mn timeout, so the regular loop is normal. But your proxy in-between would shorten this connection. I'll ping @jthabet to update the upcoming docs on that subject thanks for reporting
@cdujeu I mistaken in second option name: Here right one:
http {
keepalive_timeout 600s;
}
Btw its strange that once per 10 minutes this error will be appearing and in tray all users see " ! " sign.
i agree, i guess we should "debounce" events to make sure they do not switch the status too quick
I use this doc to setup nginx+cells: https://github.com/pydio/cells-sync/files/3784576/nginx_reverse_proxy.pdf
Once per 1-5 minutes i see in log this: