-
Notifications
You must be signed in to change notification settings - Fork 2.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
rook Failed to connect to the external object storage of the cluster #14211
Comments
What is your rook cluster manifest configuration ? |
cluster.yaml
operator.yaml
my k8s cluster, I also set node labels and storage node stains to prevent other Pods from scheduling to this node, and only tolerable Pods are allowed to schedule to this node
I also tried to create rgw within the cluster, because I set the taint, I have no way to deploy rook-ceph-rgw-my-store-a-b9b898c4d-hbmbd in the blemish In cluster.yaml and operator.yaml, I did not see signs of type CSI_CEPHFS_PLUGIN_TOLERATIONS tolerating rgw
I also tried to define cr using tolerations, but this field was not implemented in the estimation code, and an error occurred
yaml
|
The thing about your config is , you have rook cluster that is running inside the cluster, if you want to use external ceph cluster you need to create a rook-cluster with external if you want to have 2 ceph cluster one inside and another cluster outside of your k8s , then you need to create antoher rook cluster with external flag true |
You need to follow the full documentation, https://rook.io/docs/rook/latest-release/CRDs/Cluster/external-cluster/external-cluster/
`) First run python script with --rgw-endpoint flag https://rook.io/docs/rook/latest-release/CRDs/Cluster/external-cluster/external-cluster/#1-create-all-users-and-keys If you have the secret created then you can create the external object-store.yaml Plus you have to use the cluster-external.yaml |
Is it necessary to create another rook cluster to connect external ceph clusters as ehsan310 said? Why can't you use a rook cluster to connect multiple external clusters and just create an external rgw connection? Why can't rook be managed in a unified manner? A rook management itself can also manage multiple external clusters, but just create a few more CRS, which looks simpler to manage |
What configuration should I refer to for interconnection? , there is no reference link. |
you are use rook-oprator to create multiple rook-cluster. |
ok, it seems that I still need to experiment a few more times. Currently, I am not familiar with rook. Besides, why does CephObjectStore not support tolerance |
@kubecto use this document to connect rook to other rook clusters externally https://rook.io/docs/rook/latest-release/CRDs/Cluster/external-cluster/external-cluster/#exporting-rook-to-another-cluster |
Please be advised that Rook v1.10 is unsupported. Rook only provides support for the most recent 2 versions, which are v1.13 and v1.14 currently. Especially if you are new to Rook and experimenting, I would highly recommend that you begin with v1.14. |
From the original issue, it looks like you are also connecting to Ceph v14 nautilus, which has been out of support for some time. Are you able to upgrade Ceph soon, as well as to a newer version of Rook? See also Rook's official release cycle doc. We will still attempt to answer questions, but it is more difficult when such older versions are in use. |
OK, problem solved. I'll shut it down |
Is this a bug report or feature request?
Deviation from expected behavior:
According to https://rook.io/docs/rook/v1.10/Storage-Configuration/Object-Storage-RGW/object-storage/#connect-to-an-external-object-store here configuration, I tried to connect to a ceph cluster
This is my connection configuration
If I try to follow the documentation
The following error occurs, this field is not supported, so I removed it
Also, I can telnet to this address to make sure it's working.
But after I try to run it, the operator container doesn't log anything important, it just cycles through the following error
When I try to delete it, it gets stuck
When I restart a terminal, I find an error message from the operator
My own feeling is that if you want to connect a ceph cluster, you need accesskey and secretAccessKey under normal circumstances, but the rook document does not specify this, I seem to think that this is the problem here, but how to solve it? In addition to
Is the field unsupported bug affecting connections to external clusters? Can you help me explain?
I tried to create a secret and delete po operator connecting to rgw. Currently, there is no output in the log, only check
operator log
Expected behavior:
I hope that I can connect to external clusters normally, and the document needs to explain which versions of ceph clusters are supported to connect to, in addition, how to declare the connection authentication information or enter it in the CR field. At present, I am not sure how to connect to external clusters
How to reproduce it (minimal and precise):
Try using rook-1.10.12, when I am using k8s-1.28.6 cluster, and the external cluster is ceph version 14.2.22 from nautilus version, I feel like I need a key to connect to the external cluster in order to connect properly, However, these logs did not come out and the CR was deleted by the operator, nor could it be deleted normally. I was confused
Cluster Status to submit:
Environment:
uname -a
):rook version
inside of a Rook Pod):ceph -v
):kubectl version
):ceph health
in the Rook Ceph toolbox):The text was updated successfully, but these errors were encountered: