Helm Fails To Deploy Chart As Written

[CrosleyZack]

We are having issues deploying this helm chart to our local infrastructure. Following the deployment guide we have edited the values.yaml to:

1. Have a valid licenseKey, encryptionKey, and jwtSecret
2. Set the postgres host, port, db, user, password to our local postgres instance
3. Set tag to 3.114.3-stable
4. Set ingress.host and ingress.tls
5. Updated resources.limits to match our dev environment limits (cpu: 2000m, memory: 84000Ki)
6. set workflows.enabled and codeExecutor.enabled to true

After running helm install retool retool/retool -f values.yaml -n <namespace> we get the replica set created:

➜ kubectl get rs -n <namespace> -o custom-columns=":metadata.name" | grep retool
retool-55b87d976f
retool-code-executor-5f564f76d
retool-jobs-runner-54984cf5fc
retool-workflow-backend-bd884685f
retool-workflow-worker-c5c95c9b

With a created pod:

➜ kubectl describe rs -n <namepsace> retool-55b87d976f
Name:           retool-55b87d976f
...
Events:
  Type    Reason            Age   From                   Message
  ----    ------            ----  ----                   -------
  Normal  SuccessfulCreate  58m   replicaset-controller  Created pod: retool-55b87d976f-gfkbv

originally we encountered an issue with sed attempting to run on ./dist*/**/*.html

k logs -f -n <namespace> retool-7cfdc5c698-pcwqr
wait-for-it.sh: waiting for retool-postgresql:5432 without a timeout
wait-for-it.sh: retool-postgresql:5432 is available after 0 seconds
sed: can't read ./dist*/**/*.html: No such file or directory
Java dbconnector started on port 3007 with pid 36
./docker_scripts/start_api.sh: line 87:    37 Killed                  node --openssl-legacy-provider --no-experimental-fetch bundle/main.js

but failing from having no matching files:

retool_user@retool-cf86967c8-2srk6:/retool_backend$ ls -l ./dist*/**/*.html
ls: cannot access './dist*/**/*.html': No such file or directory

retool_user@retool-cf86967c8-2srk6:/retool_backend$ ls -l ./dist*/*.html
-rwxrwxrwx 1 retool_user retool_user_group 4912 Dec 16 23:47 ./dist/braintree-card-form.html
-rwxrwxrwx 1 retool_user retool_user_group  424 Dec 16 23:47 ./dist/custom-component-collections.html
-rwxrwxrwx 1 retool_user retool_user_group  181 Dec 16 23:47 ./dist/custom-components-doctype.html
-rwxrwxrwx 1 retool_user retool_user_group  255 Dec 16 23:47 ./dist/custom-components-resize.html
-rwxrwxrwx 1 retool_user retool_user_group  166 Dec 16 23:47 ./dist/custom-components.html
-rwxrwxrwx 1 retool_user retool_user_group 6730 Dec 16 23:47 ./dist/index.html
-rwxrwxrwx 1 retool_user retool_user_group 3340 Dec 16 23:47 ./dist/runtime.YjU0NTRk.html
-rwxrwxrwx 1 retool_user retool_user_group  301 Dec 16 23:47 ./dist/sandbox.html
-rwxrwxrwx 1 retool_user retool_user_group 3480 Dec 16 23:47 ./dist/stripe-card-form.html
-rwxrwxrwx 1 retool_user retool_user_group 5590 Dec 16 23:47 ./dist_webpack/braintree-card-form.html
-rwxrwxrwx 1 retool_user retool_user_group  521 Dec 16 23:47 ./dist_webpack/custom-component-collections.html
-rwxrwxrwx 1 retool_user retool_user_group  508 Dec 16 23:47 ./dist_webpack/custom-components-doctype.html
-rwxrwxrwx 1 retool_user retool_user_group  582 Dec 16 23:47 ./dist_webpack/custom-components-resize.html
-rwxrwxrwx 1 retool_user retool_user_group  493 Dec 16 23:47 ./dist_webpack/custom-components.html
-rwxrwxrwx 1 retool_user retool_user_group 7254 Dec 16 23:47 ./dist_webpack/index.html
-rwxrwxrwx 1 retool_user retool_user_group 6107 Dec 16 23:47 ./dist_webpack/runtime.cf9e15c4.html
-rwxrwxrwx 1 retool_user retool_user_group 1049 Dec 16 23:47 ./dist_webpack/sandbox.html
-rwxrwxrwx 1 retool_user retool_user_group 4068 Dec 16 23:47 ./dist_webpack/stripe-card-form.html

We attempted to resolve the issue by adding a matching html file:

$ k get deploy -n <namespace> retool -o json | jq '.spec.template.spec.containers[0].args'
[
  "bash",
  "-c",
  "chmod -R +x ./docker_scripts; sync; ./docker_scripts/wait-for-it.sh -t 0 \"retool-postgresql\":\"5432\"; sleep 1500; mkdir dist/stuff; touch dist/stuff/test.html; ./docker_scripts/start_api.sh"
]

but we now see the containing crash looping:

➜ kubectl logs -n <namepsace> retool-55b87d976f-gfkbv
wait-for-it.sh: waiting for retool-postgresql:5432 without a timeout
wait-for-it.sh: retool-postgresql:5432 is available after 0 seconds

➜ k exec -it -n <namespace> retool-55b87d976f-gfkbv -- /bin/sh
error: Internal error occurred: unable to upgrade connection: container not found ("main-backend")

➜ kubectl describe pods -n <namespace> retool-55b87d976f-gfkbv
Name:             retool-55b87d976f-gfkbv
...
Events:
  Type     Reason     Age                   From     Message
  ----     ------     ----                  ----     -------
  Warning  Unhealthy  46m (x25 over 65m)    kubelet  Liveness probe failed: Get "http://10.49.16.164:3000/api/checkHealth": dial tcp 10.49.16.164:3000: connect: connection refused
  Warning  Unhealthy  21m (x110 over 65m)   kubelet  Readiness probe failed: Get "http://10.49.16.164:3000/api/checkHealth": dial tcp 10.49.16.164:3000: connect: connection refused
  Normal   Pulled     6m26s (x19 over 66m)  kubelet  Container image "tryretool/backend:3.114.3-stable" already present on machine
  Warning  BackOff    83s (x180 over 56m)   kubelet  Back-off restarting failed container main-backend in pod retool-55b87d976f-gfkbv_accounts(bb5d06ff-798a-4052-884d-9af60e117bbf)