-
Notifications
You must be signed in to change notification settings - Fork 2.7k
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
Pvc status is pending #4251
Comments
@xiaoqge please check csi-cephfs-provisioner logs for more details on why its failing |
@Madhu-1 Thank you very much, I want to try to deploy from a new installation. In the process of deleting pvc, I checked the csi-cephfs-provisione log as you said, and I see the following error message:
Is such a mistake missing secrets? Is there a document for reference to fix? |
rook should create this secret. can you please provide the operator logs to check whats happening |
and also provide what rook version you are using |
@Madhu-1 I am using the current rook-ceph master branch,I tried to specify a namespace to create pvc, but the error log still exists。
The following is the error log
Thank you again。 |
@xiaoqge please provide rook operator pods logs |
@Madhu-1 This is some of the latest rook operator pods logs
|
@xiaoqge I don't see any error logs related to secret creation, make sure your rook image is up to date. if yes, can you please provide the list of all secrets in the rook-ceph namespace @leseb PTAL |
You're clearly not using rook master, the log message |
@leseb Thanks, in order to ensure that my operation is no problem, I tried to download the latest version of rook-1.1.6 to reinstall the deployment, but the problem still exists. Below is the specific information:
How can I solve this problem, thank you for your help. |
the logs say storage-class not found please check
|
Thanks, I posted the error log that created the storageclass. Below is the error message that is now displayed.
Below is the csi-provisioner error log:
|
please provide cephfs-plugin container logs, am suspecting some commands are struck, if possible exec into the cephfs-plugin container and check current running process |
@Madhu-1 Thank you
Below is the node log
|
i dont see any logs for command stuck ,can you try "ps -aux|grep rados" command in cephfs-plugin container |
@Madhu-1 Thank you,Below is the result of executing the "ps -aux|grep rados" command in the csi-cephfsplugin container of the k8s node. Did not see the process of rados
|
have you created filesystem? the above provided in #4251 (comment) is not complete because request logging starts from |
|
form logs @xiaoqge please provide ceph version |
Version is as follows
|
Same here... PVC stays in pending and throws this error:
|
@kevinfaguiar @xiaoqge please post the @xiaoqge please use the documentation for the Rook version you are running instead of |
@galexrt Thank you I use kubernetes 1.13.7,CentOS Linux release 7.5.1804, The rook version I am using 1.1.6
I compared https://rook.io/docs/rook/v1.1/ceph-filesystem.html and the content is no different from the master. |
My Rook version is:
My Kubernetes version is (I'm using AWS EKS):
Also, for the running pods. there is some that are not running in a pending state... could be it affecting PVC provisioning?
As for events on rock-ceph-osd-{ID} pods, the ones in pending state:
How much CPU + Memory is needed for Rook? (There are some minimum values, right? Like 2Gi of memory...) For cluster creation I'm using cluster-on-pvc, which is basically this file: https://github.com/rook/rook/blob/release-1.1/cluster/examples/kubernetes/ceph/cluster-on-pvc.yaml |
@Madhu-1 Is there any good solution? |
@kevinfaguiar Need to let your osd run normally, otherwise pvc will wait, your problem is different from what I encountered, yours is not a BUG |
What could be the cause of this problem? I can try to verify the problem, I don't have any ideas at the moment. thank you very much |
@xiaoqge looks like its cephfs core issue can you provide |
Thank You,Enter the k8s_csi-cephfsplugin_csi-cephfsplugin-provisioner container to execute the ps results as follows:
|
@xiaoqge thanks for sharing the output, I can confirm fs subvolume create commands are stuck are you able to create subvolume from toolbox pod? if yes can you please check the same from the cephfs-plugin container. |
Thank you, the conclusion is correct Below is the result of my command
Below is the log of cephfs-plugin. I didn't see my cephfs created fs volume log.
|
@Madhu-1 Is there any way I can rule out this problem, is my use incorrect or bug? |
Got the same issue when use release-1.1. Any way to fix it? |
@joscollin take a look |
I have deployed rook-ceph v1.1 on CentOS7 with mainline kernel 5.3.11 And I cannot successfully create PVC/PV on csi-cephfs Please find the
|
For your reference - There are cephfs secrets present under rook-ceph namespace
|
@xiaoqge the root cause is that your Ceph cluster isn't healthy You need to fix that,
Please fix the too few PGs per OSD issue Thanks @ShyamsundarR for this analysis. |
@satwikk, please share your ceph cluster status and ceph versions from within the rook tool box pod
|
The problem cannot be solved temporarily, it has been abandoned, thank you for your help。 |
@xiaoqge I have the same issue, it turned out that it is caused by #5938 |
I am using rook 1.1.6, kubernetes 1.13.7,CentOS Linux release 7.5.1804。I use the cephfs status to display pending,
Below is my specific configuration information:
Filesystem configuration information
Storageclass configuration information
I tried to create a Pvc and there was a pending state.Below is my pvc information
My ceph cluster status
Where am I wrong? What method can I use to find the problem? thank you very much.
The text was updated successfully, but these errors were encountered: