-
Notifications
You must be signed in to change notification settings - Fork 71
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
libvirt: support s390x cluster #1597
libvirt: support s390x cluster #1597
Conversation
15f5bda
to
833a324
Compare
6a46210
to
8eb704a
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Running through this I've hit:
time="2023-11-23T11:01:55Z" level=info msg="Cluster provisioning"
F1123 11:01:55.014896 18134 env.go:369] Setup failure: Storage pool 'default' not found. It should be created beforehand
FAIL github.com/confidential-containers/cloud-api-adaptor/test/e2e 0.028s
which might be related to the comment I added?
When I manually ran:
before |
I hit another error:
so I'm not sure if I have missed an package install? |
Please check the "-P arch=s390x" is added to the |
8eb704a
to
8e06a4b
Compare
I tested with kcli
And install the required packages with:
|
my env:
|
Hey DaLi, I'm still hitting an issue with the storage pool after creating a brand new zVSI to test this on.
The error I get is:
when I manually check it looks like it hasn't been created:
When I ran |
Running kcli_cluster.sh create manually I'm still hitting the same error and it's passing in the arguments correectly:
|
@liudalibj - I've found the reason for this. That error and storage pool check is done as part of CreateVPC, which runs before CreateCluster where your script changes would be run. I've dicussed with Wainer whether it's more appropriate for the pool storage checks to move into the podvm upload stage, so that might be a solution to this issue? |
Ah, got it, I use In your case, you didn't run the |
hi @liudalibj @stevenhorsman ! Notice that the storage pool is not necessarily the same for the e2e tests (i.e. the pool where the podvm will be pushed to) and kcli (i.e. the pool where kcli will built the nodes images). On |
And, the |
55db034
to
28aa972
Compare
Good idea, I created a new script, config_libvirt.sh, users/workflows can use it to config libvirt and kcli. |
28aa972
to
58d6235
Compare
I reproduced this issue on Ubuntu22.04 s390x vsi, so it seems that the QEMU binary do not support |
yeah |
- 'ps2 is not supported by this QEMU binary' confidential-containers/cloud-api-adaptor#1597 (comment) Signed-off-by: Da Li Liu <liudali@cn.ibm.com>
Create a fix pr karmab/kcli#623 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've re-tried on a 20.04 zVSI and have few suggestions. Manually editing the scripts, I'm able to get futher in the process, by the e2e tests are failing at the operator install phase, with:
# kubectl get pods -A
NAMESPACE NAME READY STATUS RESTARTS AGE
confidential-containers-system cc-operator-controller-manager-857f844f7d-97n7g 2/2 Running 1 (4m22s ago) 5m24s
confidential-containers-system cc-operator-daemon-install-t4v7p 0/1 ContainerCreating 0 4m36s
The describe event on that image just seems to suggest that it is taking >5mins to pull:
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 4m58s default-scheduler Successfully assigned confidential-containers-system/cc-operator-daemon-install-t4v7p to peer-pods-worker-0
Warning FailedCreatePodSandBox 4m58s kubelet Failed to create pod sandbox: rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial unix /var/run/containerd/containerd.sock: connect: no such file or directory"
Normal Pulling 4m43s kubelet Pulling image "quay.io/confidential-containers/runtime-payload-ci:kata-containers-8de1f8e19f858134ba455a7c04edcb21d8bcf6b1"
I'm hoping that's a network issue with quay or something and I'll try again later
The separate config script is really helpful to me when doing local tests @wainersm are you ok with this approach? |
- 'ps2 is not supported by this QEMU binary' confidential-containers/cloud-api-adaptor#1597 (comment) Signed-off-by: Da Li Liu <liudali@cn.ibm.com>
Hey DaLi, just to give you an update on this. I had the same problem this morning on Ubuntu 22.04, but then had a thought that the profile I was using might be the issue. I was running it on a cz2-4x8 zVSI, which is the same size I use for x86, but I checked and saw you were using bz2-8x32, so I've retried with that and got some e2e tests running now... |
fe27b47
to
72001c2
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'd like to see @wainersm review this before it's merged, but I've tried it out on a 8 vCPU, 32GB RAM s390x VM and all the tests pass. Thanks for all the great work @liudalibj!
Hi @liudalibj ! First, thanks for accepting my suggestion! I tried it out locally on my x86_64 machine, running with I got this error:
That I fixed with:
Having that the e2e tests passed on my machine \o/ |
- create s390x cluster with libvirt - show e2e-test result for s390x libvirt cluster Signed-off-by: Da Li Liu <liudali@cn.ibm.com> follow the review comment, use a new script to config libvirt
72001c2
to
fb2ec83
Compare
Thanks @wainersm I updated the script base on your comment. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
@wainersm do you have any new comments? If there is none, I would like merge this pr tomorrow. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @liudalibj !
fixes #1598