howdy! i’ve deployed Flyte in AWS manually in EKS per
the docs. things look good from a deployment perspective - the deployment, pods, services, and ingresses are up. i can hit the console web UI at the expected hostname. two things i havent done are get our SSL certs working on the ingress yet or set an authentication method. but i wanted to get the
flytectl
config set and be able to interact with the cluster via the command line with a minimum configuration before doing those two things.
my minimal config looks like this
admin:
endpoint: dns:///admin.flyte.us3.predictap.com
insecure: true
my
flytectl version
output is this
{
"App": "flytectl",
"Build": "62b86f6",
"Version": "0.6.7",
"BuildTime": "2022-08-11 11:33:31.781304 -0500 CDT m=+0.021446651"
}%
i think my config is good enough, i ran
flytectl config validate
chrisantenesse@Chriss-MacBook-Pro-2 ~ % flytectl config validate
Couldn't find a config file.
Validated config file successfully.
but when i do something like
flytectl get projects
, i’m getting this
{"json":{},"level":"error","msg":"failed to initialize token source provider. Err: failed to fetch auth metadata. Error: rpc error: code = Unavailable desc = connection closed","ts":"2022-08-11T11:35:12-05:00"}
{"json":{},"level":"warning","msg":"Starting an unauthenticated client because: can't create authenticated channel without a TokenSourceProvider","ts":"2022-08-11T11:35:12-05:00"}
{"json":{},"level":"info","msg":"Initialized Admin client","ts":"2022-08-11T11:35:12-05:00"}
Error: rpc error: code = Unavailable desc = connection closed
{"json":{},"level":"error","msg":"rpc error: code = Unavailable desc = connection closed","ts":"2022-08-11T11:35:12-05:00"}
i jumped on both flyteadmin pods shell and was able to confirm that traffic coming through the ingress was actually hitting the pods. basically ran
watch -n1 netstat -anp
and watched a connection get established. i also ran
tcpdump
locally and watched my local machine make the outbound request, etc. so i dont think this is due to a misconfiguration in the k8s side of things.
i ran a
kubectl logs …
on both pods, but never saw entries from the flyteadmin server indicating that something happened (good or bad).
im new to the flyte world and appreciate the help in advance!
