Make slaves run in memory mode NONE about helmchart HOT 2 CLOSED

Something is wrong with the liveness probe. I did an upgrade and two of the three slaves I have

[christopher@chris-msi ~]$ kubectl logs netdata-slave-pdclb
...
2019-04-29 18:25:27: netdata ERROR : MAIN : Failed to read machine GUID from '/var/lib/netdata/registry/netdata.public.unique.id'
2019-04-29 18:25:27: netdata FATAL : MAIN :Cannot create unique machine id file '/var/lib/netdata/registry/netdata.public.unique.id'. Please fix this. # : Invalid argument
...

After that, netdata exited. It looks like something went wrong with the preStart command that creates this file, probably a race condition. When I deleted the pod and it was recreated, it worked fine.

I'll try to replicate the issue with preStart, but the worrisome thing is that the two problematic pods were showing as running, instead of error. I'll try to get on this tomorrow, it's not good.

from helmchart.

I saw that we didn't have readiness/liveness probes in the daemonset (no web server running anyway), so I added them. It takes much longer for the slaves to show up as ready now, but I expect it to work better. I will close this issue with PR #19 and create a new issue to check on the apparent preStart race condition.

from helmchart.