Logs sometimes silently stop loading
Closed this issue · 2 comments
JayH5 commented
Mike reported logs that were working but then suddenly stopped. When I tried to load the logs I got nothing. Logs still working fine in Mesos UI.
Not really sure what causes this and the best I have are these logs from Nginx (error.log) in the MC2 containers:
2016/04/15 14:58:56 [error] 43#0: *40358 upstream timed out (110: Connection timed out) while reading response header from upstream, client: 172.17.0.1, server: , request: "GET /base/logs/84/b435432f-0313-11e6-9c88-ba9f87ec83a0/stderr/?n=1 HTTP/1.0", upstream: "http://0.0.0.0:8000/base/logs/84/b435432f-0313-11e6-9c88-ba9f87ec83a0/stderr/?n=1", host: "qa-mesos-mc.za.prk-host.net", referrer: "http://qa-mesos-mc.za.prk-host.net/base/logs/84/"
2016/04/15 15:00:47 [error] 43#0: *40368 upstream prematurely closed connection while reading response header from upstream, client: 172.17.0.1, server: , request: "GET /base/logs/84/b435432f-0313-11e6-9c88-ba9f87ec83a0/stdout/?n=1 HTTP/1.0", upstream: "http://0.0.0.0:8000/base/logs/84/b435432f-0313-11e6-9c88-ba9f87ec83a0/stdout/?n=1", host: "qa-mesos-mc.za.prk-host.net", referrer: "http://qa-mesos-mc.za.prk-host.net/base/logs/84/"
Nothing to be seen in MC2 logs. Nginx access logs report 502/504 response codes.