0

No matter if I go with default configuration or custom (reducing cpus, because I'm on trial) my maraton-lb can't start and fails on health check non stop. What am I doing wrong?

Log:

[ALERT] 184/143843 (321) : Starting frontend marathon_http_in: cannot bind socket [0.0.0.0:80]
[ALERT] 184/143843 (321) : sendmsg logger #1 failed: No such file or directory (errno=2)
[ALERT] 184/143843 (321) : sendmsg logger #2 failed: No such file or directory (errno=2)
[ALERT] 184/143843 (321) : sendmsg logger #1 failed: No such file or directory (errno=2)
[ALERT] 184/143843 (321) : sendmsg logger #2 failed: No such file or directory (errno=2)
Received task health update, healthy: false
Received killTask for task marathon-lb.cbfaf218-7f97-11e8-aef1-70b3d5800001

marathon-json:

{
  "marathon-lb": {
      "instances": 1.0, "mem": 1048.0, "cpus": 1
  }
}

dcos package install --options=marathon-config.json marathon-lb

FYI I tried a loooot of configs. Azure docs, marathon docs, mesos docs and all sorts of different variations. Always the same result.

Afterlook
  • 65
  • 9

1 Answers1

0

Do you have something else already bound to port 80 ? Marathon-lb is usually deployed to a public agent, so the config is usually longer than the config you've posted - have you tried with the default config ? By default it also needs to bind to 80 and 443, together with the unprivileged ports it uses - that error looks like it can't bind wherever it's been deployed.