All Posts

Hi, thank you for the reply. I am aware they could be selected. My questions are: Has it been tested by VMware? (I do not want to spend time to create a lab to test it, if this scenario has not even been tested by VMware) Is it supported? (or more exactly will it be a supported configuration when CSE 4 will be in GA.)

I was trying to generate a template but it didn't work testtenant@desktop:~$ clusterctl generate cluster test Error: failed to read "cluster-template.yaml" from provider's repository "infrastructure-vcd": failed to get GitHub release v1.0.0: failed to read release "v1.0.0": GET https://api.github.com/repos/vmware/cluster-api-provider-cloud-director/releases/tags/v1.0.0: 404 Not Found [] Clusterctl support for Cloud Director is enabled only on clusterctl 1.2. However, our latest CAPVCD (Cluster API provider for Cloud Director) is still configured to work with Core CAPI 1.1.3 and Cluster 1.1.3 only. We will very soon qualify CAPVCD with CAPI 1.2. We will keep you posted on this.

Hi  If your goal is to automate everything (create, resize, upgrade, delete) via VCD API, that is doable, but you must take extra care to craft the payload correctly.  If your goal is to create the cluster using CSE 4.0 UI and then use VCD API to manage it further, that is also doable.  If your goal is to create the cluster using CSE 4.0 UI and use kubectl (Cluster API) to manage it further, that is not straightforward and we do not recommend this. There is a way to get this option also working. If this is what you want, let us know. We can provide the details. Which one of the above is your requirement?

This will be available as part of CSE4. The `controlPlaneIP` field of the capi yaml can be edited if capvcd is to be used directly. However that will be part of CAPVCD 1.0.0. (The change is in the main branch.)

I have done some tests regarding this topic with a cluster created with 3 master. If one control plane node is shutdown from vCenter, "get pods -A" continue to work. (As expected) If two control plane nodes are shutdown, "get pods -A" doesn't work anymore (Expected) After restarting one of the control plane node "get pods -A" works again, (Expected) So the basic functionality of a multi control plane nodes is working. One issue is that no errors are reported in the events or in status of the cluster from CSE plugin. (Status is "ready") The only thing visible is at load balancer level  which shows that some endpoints are down and VAPP that is noticing some VMs down. Would it be possible to add some kinds of "health" in the CSE plugin? (like all control planes node up and running / worker nodes up and running, load balancer associated to management IP deployed etc) Second issue, I have deleted on purpose one of the control plane VM. As mentioned above no information are reported from the CSE plugin, it still show "3 nodes". It doesn't recreate the missing node (no "auto-heal" , which would be the best) Is there a procedure on how to replace a failed node in such case?

Yes, it does, but only If I login to each site separately. If I login to rs-bg-2 and switch to rs-bg-1 site from the "Data centers" page of tenant portal, then I get "Error: Failed to fetch Kubernetes clusters" because of "401 Unauthorized" request to /api/session and the rezulting "CSE UI: Error fetching sites". Since this is evident even on CSE 3, should I open SR and debug it there and give VMware with screen sharing session so someone can better describe problem and collect what is going on?

Hi @marko_uskokovic  Sorry for the delay; my message did not go through.   Would you please log into each site and go to the container plugin on each site? Does each site show clusters successfully?   Thank you

Hi, I do not have full details but from what i understood: NSX ALB communicate with vCenter using a "vCenter account" dedicated for this purpose. (This is part of "create NSX-T Cloud) in vcenter. So it seems somehow that NSX-ALB was not able to communicate with vCenter anymore. So maybe password has been modified or something like this. Note: I am maybe mistaken an issue was with account connecting to NSX-Manager but the concept is the same, issue with credentials used with NSX-T cloud) After fixing credentials the deployment was successful. Summary: The issue was not related to Tanzu/CSE but the underlying NSX-ALB infrastructure. Unfortunately it is not easy to pinpoint the origin when looking at error at Tanzu/CSE level. Therefore, the feature requests of adding "pre-requisite" check and/or a wizard showing the progression of a cluster deployment step by step. (Showing the steps completed, current step, and next steps.) In this way it would be easier to pinpoint the origin of such issue if one step is stuck.

Hi, any update on this request? I was trying to generate a template but it didn't work testtenant@desktop:~$ clusterctl generate cluster test Error: failed to read "cluster-template.yaml" from provider's repository "infrastructure-vcd": failed to get GitHub release v1.0.0: failed to read release "v1.0.0": GET https://api.github.com/repos/vmware/cluster-api-provider-cloud-director/releases/tags/v1.0.0: 404 Not Found [] Or is clusterctl generate still not supported for vcloud director? Like in this link clusterctl generate command doesn't support the generation of CAPI yaml for Cloud Director; Follow the guidelines provided below configure the CAPI Yaml file cluster-api-provider-cloud-director/WORKLOAD_CLUSTER.md at main · vmware/cluster-api-provider-cloud-director · GitHub

I don't see any specific error, because it opens in new tab without inspect element so I can't track http requests. Will report if I find some error. When I login to rs-bg-2-ec site and want to go to another site by SSO I just click on site name in "Data Centers" -> "Virtual Data Center" like in picture attached  

Hi @marko_uskokovic, Thank you for your feedback. We are looking into this. When SSO login fails after clicking the cse4-rs-bg-1-test-1 link on the rs-bg-2-ec, do you see any specific error? Would you please also elaborate on when you log into the rs-bg-2-ec site and go to another site by SSO, what specific steps do you do to get to the other site?   Thank you,   Timothy

>Unfortunately I couldn't easily get into the VM to investigate further - the password found by editing the guest properties in vCD does >not work and nor does connecting from the machine that has the SSH key. Unless a different user (not root) should be used? I have also >reset the password by booting to single-user mode, but it appears to be overwritten by cloud-init once the VM boots again. These are known issues that are getting resolved in GA release > When I restarted the CSE service it deleted the VMs New flag to keep the vm(s) on failure is part of GA. But did the retry on cluster creation happen ?