miq orchestrator container doesn't propagate right environment variables to workers

[p-v-a]

I'm running jansa-1 version on vanilla k8s 1.17.2. I have several issues with miq orchestrator container either not passing required information into worker containers or MIQ ignores it

1. Secrets names are hardcoded - postgresql-secrets, kafka-secrets, app-secrets are hardcoded in MIQ and as far as I can see there is no way to overwrite it see - databaseSecret not propagated to the workers manageiq-pods#601

2. Even with correct secrets worker pods in a crashloop state
   NAME READY STATUS RESTARTS AGE
   1-event-handler-6cc5bfff7c-7rj96 1/1 Running 5 4m24s
   1-generic-54855c57fd-c76sx 0/1 CrashLoopBackOff 3 3m46s
   1-generic-54855c57fd-mv2fp 0/1 CrashLoopBackOff 4 3m46s
   1-priority-5994c67b7d-s7jrs 0/1 CrashLoopBackOff 4 4m24s
   1-priority-5994c67b7d-zpscq 1/1 Running 4 4m24s
   1-remote-console-745d976499-8krqg 0/1 Error 5 4m25s
   1-reporting-6fd59749c6-4l9k5 0/1 CrashLoopBackOff 4 3m27s
   1-reporting-6fd59749c6-zmjsc 0/1 CrashLoopBackOff 3 3m27s
   1-schedule-5fb84c567b-264f2 0/1 CrashLoopBackOff 4 3m48s
   1-ui-7d9dcd76db-zfdxr 0/1 Running 1 27s
   1-web-service-677b86f9f7-w28l9 0/1 CrashLoopBackOff 1 100s
   miq-dev01-memcached-667dfb8767-xzgn2 1/1 Running 0 150m
   miq-dev01-orchestrator-58dc4b567-gg755 1/1 Running 0 123m
   miq-dev01-postgresql-deployment-7497f5d5c7-rn5z4 1/1 Running 0 161m

with the errors

{"@timestamp":"2020-09-23T05:26:03.425134 ","hostname":"1-ui-7d9dcd76db-l8d7z","pid":8,"tid":"2ad73e4f195c","level":"info","message":"MIQ(SessionStore) Using session_store: ActionDispatch::Session::MemCacheStore"}
{"@timestamp":"2020-09-23T05:26:03.885417 ","hostname":"1-ui-7d9dcd76db-l8d7z","pid":8,"tid":"2ad73e4f195c","level":"debug","message":"Dalli::Server#connect 127.0.0.1:11211"}
{"@timestamp":"2020-09-23T05:26:03.886121 ","hostname":"1-ui-7d9dcd76db-l8d7z","pid":8,"tid":"2ad73e4f195c","level":"warning","message":"127.0.0.1:11211 failed (count: 0) Errno::ECONNREFUSED: Connection refused - connect(2) for "127.0.0.1" port 11211"}

Looks like it defaults to local memcached if no environment variables are provided. I think orchestrator passes MEMCACHED_SERVER and MEMCACHED_SERVICE_NAME into workers. Yet orchestrator itself expects memcached connection details in MEMCACHED_SERVICE_HOST and MEMCACHED_SERVICE_PORT.
Even if I add MEMCACHED_SERVER and MEMCACHED_SERVICE_NAME to the pod definition it looks like it never used, I see same error in the logs.

[p-v-a]

All worked pods were started successfully after I patched worker deployment definitions by adding MEMCACHED_SERVER and MEMCACHED_SERVICE_NAME variables.

[p-v-a]

Sorry I confused myself. I have added MEMCACHED_SERVICE_HOST and MEMCACHED_SERVICE_PORT, not MEMCACHED_SERVER and MEMCACHED_SERVICE_NAME variables I mentioned above.

[chessbyte]

@bdunne @abellotti is this still an issue in Kasparov release?