cloud-bulldozer/e2e-benchmarking

Error from server (AlreadyExists): configmaps "kube-burner-[UUID]" already exists

dry923 opened this issue · 4 comments

I ran two sets of node-density light tests back to back. I did not have CLEANUP_WHEN_FINISHED set to true however CLEANUP is set to true. On the second run I get an error that the configmap for the test already exists. The UUID variable was never unset so it attempted to reuse the same one. The old namespace and benchmark were deleted. We should add the configmap cleanup to the CLEANUP task as well to ensure we properly clean everything up.

# ./run_nodedensity_test_fromgit.sh                                                                            
Collecting ripsaw-cli                                                   
...
Mon 21 Feb 2022 12:17:50 PM UTC Number of pods to deploy on nodes: 6318                                                                   [0/1905]
Mon 21 Feb 2022 12:17:50 PM UTC Labeling 27 worker nodes with node-density=enabled
node/dry-gcp-27-hxw5p-worker-a-248c6.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-a-7lkd5.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-a-7qm9n.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-a-c5j4h.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-a-cn4p7.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-a-hbnwc.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-a-jxgfp.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-a-rh4xw.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-a-z4wvp.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-54zdn.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-5cvnc.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-7sd5p.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-cpp94.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-l9sck.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-nd9zm.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-sdz8f.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-tlzmw.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-b-xlg24.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-2n4zf.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-jmztx.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-k86vk.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-krpfb.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-m87qp.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-mqth7.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-t458g.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-wbnvn.c.openshift-perfscale.internal labeled
node/dry-gcp-27-hxw5p-worker-c-ztss7.c.openshift-perfscale.internal labeled
cp: cannot stat 'metrics-profiles/metrics-profiles/metrics.yaml': No such file or directory
Mon 21 Feb 2022 12:18:03 PM UTC Creating kube-burner configmap
Error from server (AlreadyExists): configmaps "kube-burner-cfg-c1a83fa3-90fe-49bd-b747-16fa91c1475a" already exists

@rsevilla87 thoughts?

@dry923, the configmap name is based on the UUID, hence you have to gen a new UUID or delete the old configmap

@rsevilla87 right and I 100% understand that. My point is that if CLEANUP is enabled it should remove the configmap if it exists as well

@rsevilla87 right and I 100% understand that. My point is that if CLEANUP is enabled it should remove the configmap if it exists as well

100%, Ill file a PR