Open thecodeassassin opened 7 months ago
Hi @thecodeassassin, thanks for reporting!
Could you please also provide additional information, such as:
everest-olm
namespace-v
option)I have a similar issue today, I can't seem to install it on a fresh EKS cluster:
cli
everestctl version
ProjectName: everestctl
Version: v1.0.1
FullCommit: 40b9e1c074f75b1452143102dcf062d91bb8ab93
cluster
status
k logs -f -n everest-olm olm-operator-6f7b945cbb-rjcdd
# outputs this multiple times per second forever:
time="2024-07-08T18:55:05Z" level=info msg="install strategy successful" csv=packageserver id=uGZ4N namespace=everest-olm phase=Installing strategy=deployment
and:
k logs -f packageserver-6b766dcf95-nc7lz -n everest-olm
# has grpc errors since I believe the pod above isn't running yet:
time="2024-07-08T18:53:43Z" level=info msg="updating PackageManifest based on CatalogSource changes: {everest-catalog everest-olm}" action="sync catalogsource" address="everest-catalog.everest-olm.svc:50051" name=everest-catalog namespace=everest-olm
W0708 18:54:02.897521 1 logging.go:59] [core] [Channel #1 SubChannel #2] grpc: addrConn.createTransport failed to connect to {Addr: "everest-catalog.everest-olm.svc:50051", ServerName: "everest-catalog.everest-olm.svc:50051", }. Err: connection error: desc = "transport: Error while dialing: dial tcp [fda1:d88a:dc1::b912]:50051: i/o timeout"
time="2024-07-08T18:54:02Z" level=warning msg="error getting bundle stream" action="refresh cache" err="rpc error: code = Unavailable desc = connection error: desc = \"transport: Error while dialing: dial tcp [fda1:d88a:dc1::b912]:50051: i/o timeout\"" source="{everest-catalog everest-olm}"
time="2024-07-08T18:54:02Z" level=info msg="updating PackageManifest based on CatalogSource changes: {everest-catalog everest-olm}" action="sync catalogsource" address="everest-catalog.everest-olm.svc:50051" name=everest-catalog namespace=everest-olm
time="2024-07-08T18:54:02Z" level=warning msg="error getting bundle stream" action="refresh cache" err="rpc error: code = Unavailable desc = connection error: desc = \"transport: Error while dialing: dial tcp [fda1:d88a:dc1::b912]:50051: i/o timeout\"" source="{everest-catalog everest-olm}"
time="2024-07-08T18:54:03Z" level=info msg="updating PackageManifest based on CatalogSource changes: {everest-catalog everest-olm}" action="sync catalogsource" address="everest-catalog.everest-olm.svc:50051" name=everest-catalog namespace=everest-olm
time="2024-07-08T18:54:03Z" level=warning msg="error getting bundle stream" action="refresh cache" err="rpc error: code = Unavailable desc = connection error: desc = \"transport: Error while dialing: dial tcp [fda1:d88a:dc1::b912]:50051: i/o timeout\"" source="{everest-catalog everest-olm}"
time="2024-07-08T18:54:04Z" level=info msg="updating PackageManifest based on CatalogSource changes: {everest-catalog everest-olm}" action="sync catalogsource" address="everest-catalog.everest-olm.svc:50051" name=everest-catalog namespace=everest-olm
time="2024-07-08T19:07:32Z" level=info msg="updating PackageManifest based on CatalogSource changes: {everest-catalog everest-olm}" action="sync catalogsource" address="everest-catalog.everest-olm.svc:50051" name=everest-catalog namespace=everest-olm
This is occuring within an ipv4 EKS cluster as well for me.
@fideloper from the package server logs, it seems to me that it cannot reach the catalog server. Can you check the status of the everest-catalog-xx
pod? You can also describe the catalog source object for additional clues - kubectl describe catalogsource everest-catalog -n everest-olm
I've opened up all ports, altho the error being "connection refused" sounds like nothing is listening there rather than there being a security group in the way (which would cause a timeout typically).
Some more info:
#> everestctl -v install
2024-07-09T07:19:22.943-0400 DEBUG commands/root.go:34 Debug logging enabled
? Namespaces managed by Everest [comma separated] everest,percona
? What operators do you want to install? MySQL, PostgreSQL
2024-07-09T07:19:31.661-0400 DEBUG install/install.go:552 Enabling MySQL operator {"component": "install"}
2024-07-09T07:19:31.661-0400 DEBUG install/install.go:552 Enabling PostgreSQL operator {"component": "install"}
2024-07-09T07:19:31.852-0400 DEBUG install/install.go:214 Everest latest version available: 1.0.1 {"component": "install"}
2024-07-09T07:19:31.852-0400 DEBUG install/install.go:215 Everest version information &version.MetadataVersion{state:impl.MessageState{NoUnkeyedLiterals:pragma.NoUnkeyedLiterals{}, DoNotCompare:pragma.DoNotCompare{}, DoNotCopy:pragma.DoNotCopy{}, atomicMessageInfo:(*impl.MessageInfo)(nil)}, sizeCache:0, unknownFields:[]uint8(nil), Version:"1.0.1", Recommended:map[string]string{}, Supported:map[string]string{"cli":">= 1.0.0"}} {"component": "install"}
2024-07-09T07:19:31.852-0400 INFO install/install.go:580 Installing Operator Lifecycle Manager {"component": "install"}
2024-07-09T07:19:31.902-0400 DEBUG kubernetes/kubernetes.go:430 Applying "crds/olm/crds.yaml" file {"component": "install", "component": "kubernetes"}
2024-07-09T07:19:34.821-0400 DEBUG kubernetes/kubernetes.go:430 Applying "crds/olm/olm.yaml" file {"component": "install", "component": "kubernetes"}
2024-07-09T07:19:39.779-0400 INFO install/install.go:585 OLM has been installed {"component": "install"}
2024-07-09T07:19:39.779-0400 INFO install/install.go:586 Installing Percona OLM Catalog {"component": "install"}
2024-07-09T07:19:39.780-0400 DEBUG kubernetes/kubernetes.go:398 Using catalog image docker.io/percona/everest-catalog:1.0.1 {"component": "install", "component": "kubernetes"}
## This is where it just sits/waits
I got this working, we had to open up the security groups so the control plane could talk to the nodes over port 5443
(where as I had previously setup security groups for just intra-node communication on port 5443).
This should be internal traffic within the EKS cluster so if we have to open security group ports then something isn't architected correctly on the Percona as the traffic shouldn't be routing like that. Percona should definitely take a look at this if that's the case.
Hi all,
everestctl install is stuck on this step:
verestctl install --namespaces db-prod,db-dev --operator.mongodb=true --operator.postgresql=true --operator.xtradb-cluster=true --skip-wizard 2024-04-13T14:28:16Z info install/install.go:430 Installing Operator Lifecycle Manager {"component": "install"} 2024-04-13T14:28:16Z info kubernetes/kubernetes.go:328 OLM operator is already installed {"component": "install", "component": "kubernetes"} 2024-04-13T14:28:16Z info install/install.go:435 OLM has been installed {"component": "install"} 2024-04-13T14:28:16Z info install/install.go:436 Installing Percona OLM Catalog {"component": "install"}
it fails somewhere here:
func (k *Kubernetes) InstallPerconaCatalog(ctx context.Context, version *goversion.Version) error {
But with no errors.