deis / monitor

Monitoring for Deis Workflow
https://deis.com
MIT License
22 stars 32 forks source link

InfluxDB persistence after helm install #191

Open pfeodrippe opened 7 years ago

pfeodrippe commented 7 years ago

Could I enable InfluxDB persistence after helm install deis/workflow? I read the docs, but what would happen?

jchauncey commented 7 years ago

Yes you can do that. It would create a persistent volume and bind it to the influx pod. All data currently stored in the pod would be lost though since influx would need to restart to bind the volume.

jchauncey commented 7 years ago

I think you should be able to do something like helm upgrade <release name> <chart path> --set influxdb.persistence=true

pfeodrippe commented 7 years ago

I've did what you said, but it's taking a long time to terminate the old pods and Grafana is not showing anything. Maybe it should be better to delete the release and reinstall it? The influxapi IP should be the same, right? It's a Service, it should discover its pods

helm upgrade amber-bobcat deis/workflow -f ./values.yaml \ 
--set global.storage=s3,s3.accesskey=$AWS_ACCESS_KEY_ID,s3.secretkey=$AWS_SECRET_ACCESS_KEY,s3.region=us-east-1,s3.registry_bucket=registry-deis-staging,s3.database_bucket=database-deis-staging,s3.builder_bucket=builder-deis-staging,monitor.influxdb.persistence.enabled=true,monitor.grafana.persistence.enabled=true

$ kubectl -n deis get po -w
...
deis-monitor-grafana-1790189660-jsgg8    1/1       Running       0          10m
deis-monitor-grafana-541875647-nj7vm     1/1       Terminating   0          3d
deis-monitor-influxdb-1032455162-26jft   1/1       Running       0          10m
deis-monitor-influxdb-2881701064-h821m   1/1       Terminating   0          3d
...

$ kubectl -n deis get services
...
deis-monitor-influxapi   100.64.181.59    <none>             80/TCP                                                     10d
deis-monitor-influxui    100.64.188.135   <none>             80/TCP
...
jchauncey commented 7 years ago

If you are ok deleting the release and reinstalling then yeah I would try that. Ive recently been working on the charts and I have done several no persistence to persistence upgrades and didnt notice that issue. But that was on a branch of some stuff I was testing so it may not work that way in master.