Details
-
Bug
-
Status: Done
-
High
-
Resolution: Done
-
None
-
None
-
// code placeholder
-
Yes
Description
Hi,
We have the situation when we deploy the cluster with HAProxy and then we try to update the secrets they are not correctly updated due to the following issue:
{"level":"error","ts":1606490696.770058,"logger":"controller","msg":"Reconciler error","controller":"perconaxtradbcluster-controller","name":"cluster1","namespace":"run-k8spxc-485-log-test-5990","error":"reconcileUsers: manage sys users: update Proxy users pass: update proxy users: begin transaction: dial tcp: lookup cluster1-proxysql-unready.run-k8spxc-485-log-test-5990 on 10.7.240.10:53: no such host","errorVerbose":"dial tcp: lookup cluster1-proxysql-unready.run-k8spxc-485-log-test-5990 on 10.7.240.10:53: no such host\nbegin
As you can see we are trying to update secrets for ProxySQL but we do not have it at all and that is why we have the error and the update is stuck.
To do:
- we need to check that ProxySQL was not deployed and do not try to sync/update secrets for it
- re-check our tests and cover this case too
- we need to mount secret with user credentials as a file to avoid failure of readiness/liveness probs when secret is changed