Hi there! I got an scenario that I need to restart kapacitor (because new UDF configs).
My setup its an Docker running over kubernetes. I have a pod with two containers (one of them I used to manage udf(it changes kapacitor.conf and puts to run all .py udf)).
So, the changes to kapacitor.conf I’ve made via Kubernetes API, and also after that I send an exec “kill 1” which is a kapacitord process.
Up to this point its all right, Kubernetes detect a “failure”(made on purpose with kill 1) and re-launches the container. When container starts all seem correct, but kapacitord it’s not responding to anyone (even with localhost)
Inside kapacitor container:
curl localhost:9092/kapacitor/v1/tasks
curl: (7) Failed to connect to localhost port 9092: Connection refused
Logs from kapacitor container:
msg="kapacitor starting" service=run version=1.5.2 branch=HEAD commit=3086452d00830e01d932838d8c6d1df818648ad3
msg="go version" service=run version=go1.11.2
msg="listing Kapacitor hostname" source=srv hostname=kapacitor-77bc647c96-595b6
msg="listing ClusterID and ServerID" source=srv cluster_id=8a7427cf-7837-4132-a1f5-01cf24d28769 server_id=a21444b2-8ca2-423a-95cb-ba5fae412f3d
msg="opened task master" service=kapacitor task_master=main
PS aux output:
root@kapacitor-77bc647c96-gc4nh:/# ps aux
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.0 0.1 609384 39792 ? Ssl 14:06 0:00 kapacitord
root 19 0.0 0.0 4280 712 pts/0 Ss 14:07 0:00 sh
root 29 0.0 0.0 19092 4356 pts/0 S 14:10 0:00 bash
root 292 0.0 0.0 36636 2872 pts/0 R+ 14:11 0:00 ps aux
If I wait some time kapacitor throws(and CrashLoopBack container):
run: open server: open service *udf.Service: failed to load process info for "basic": dial unix /socket/basic.sock: connect: no such file or directory
or
run: open server: open service *udf.Service: failed to load process info for "basic": dial unix /socket/basic.sock: connect: connection refused
Any ideas? Thanks in advance
