| airship-irc-bot | <sidney.shiba> Hi all, a question on VMs used by Zuul. Is there a pool of VMs already instantiated and ready to be used by Zuul or are they dynamically instantiated (clean VM)? I ask this because if the former, is there a possibility that Docker images are cached and might not be pulling latest image from quay.io if tag is the same, e.g., helm-chart-collator:*latest*? | 15:13 |
|---|---|---|
| airship-irc-bot | <sidney.shiba> Please, disregard my previous message. Found the issue, which was that treasuremap was pinning to an old tag. | 16:49 |
| airship-irc-bot | <sidney.shiba> Hello all, Zuul is failing during 36_deploy_workload.sh with snippet of error showing below. Did anybody experience such error? | 18:21 |
| airship-irc-bot | <dwalt> I am getting those in the virtual-network-cloud deployment job, but I'm not sure if it's for the same reason or not because the logs haven't been collected | 18:47 |
| airship-irc-bot | <dwalt> It looks like the ingress pods are getting evicted because there isn't enough storage on your change | 18:47 |
| airship-irc-bot | <dwalt> https://zuul.opendev.org/t/openstack/build/f1b12769d04a4c62a82f191d8b449e98/log/prima[…]ss/pods/ingress-ingress-nginx-controller-585567484-5xp5r.txt | 18:47 |
| airship-irc-bot | <sidney.shiba> Note, I am working on _*dex-aio*_, which is part of _*workload-target*_ deployment. I didn't touch the _*ingress*_ portion so not aware of what needs to be done. Who should I reach for the _*ingress*_ implementation and figure out how to fix this resource issue? | 19:34 |
| airship-irc-bot | <dwalt> @sidney.shiba if the disk pressure is causing the failure, then it is probably because we schedule the majority of our applications to the master node. Now that we have labels, we should probably start using the correct selectors anyway | 20:55 |
| airship-irc-bot | <dwalt> So we can start by removing ingress from that node | 20:55 |
| airship-irc-bot | <sidney.shiba> Those changes will require changing in the charts repo if not mistaken, which will delay the merge for https://review.opendev.org/c/airship/treasuremap/+/791835. Isn't there something that can be done to get Zuul passing without this chain of changes to make it work? | 21:08 |
| airship-irc-bot | <dwalt> Yes, we can remove other functions from that node. Like the ingress controller | 21:09 |
| airship-irc-bot | <dwalt> Is there not a way to override the `nodeSelector` in the Dex helm chart values? | 21:09 |
| airship-irc-bot | <sidney.shiba> Yes, there is. What should be the value? | 21:11 |
| airship-irc-bot | <dwalt> `node-role.kubernetes.io/worker: ""` | 21:18 |
| airship-irc-bot | <dwalt> Our worker nodes use that label now | 21:18 |
| airship-irc-bot | <sidney.shiba> I just tested and I cannot override current _nodeSelector_ but only append to it. `over_rides:` `deployment:` `dex:` `spec:` `template:` `spec:` `nodeSelector:` `node-role.kubernetes.io/worker: ""` | 21:23 |
| airship-irc-bot | <sidney.shiba> It will require to go to the charts repo and change it there. | 21:24 |
| airship-irc-bot | <sidney.shiba> ok, found a way. | 21:28 |
| airship-irc-bot | <sidney.shiba> Tested and have the dex-aio pod moved to the worker node. Hoping it will resolve the Zuul issue and pass the gates. Thanks for the help. | 21:46 |
| airship-irc-bot | <pg588x> pls review https://review.opendev.org/c/airship/treasuremap/+/794651 | 22:17 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!