| airship-irc-bot | <raliev> @sidney.shiba, what particular issues you observed? sometimes I observe this one: ```[airshipctl] 2021/05/25 18:04:18 /root/src/airshipctl/pkg/events/processor.go:84: Received error when applying resources to kubernetes: polling for status failed: etcdserver: request timed out Error events received on channel, errors are: [polling for status failed: etcdserver: request timed out]``` increasing VM cpu/memory spec solves this problem | 00:59 |
|---|---|---|
| airship-irc-bot | for me | 00:59 |
| *** uzumaki has joined #airshipit | 04:21 | |
| *** uzumaki has quit IRC | 04:32 | |
| *** avolkov has joined #airshipit | 08:19 | |
| *** avolkov has quit IRC | 10:38 | |
| *** uzumaki has joined #airshipit | 11:41 | |
| *** uzumaki has quit IRC | 12:54 | |
| airship-irc-bot | <sidney.shiba> The errors varied in an inconsistent ways. Also it was very slow. So this combination of inconsistent errors and slow made me assume that was memory swap issues, similar to what you described. The flavors available didn't give me the choice of increase memory only. Note that there was no issue when deploying on an Azure VM (Standard | 13:03 |
| airship-irc-bot | <sirishagopigiri> Hi Yeam, Requesting reviews for below PSs related to #280 airshipctl issue https://review.opendev.org/c/airship/airshipctl/+/790020 --- needs another +2 and WF +1 https://review.opendev.org/c/airship/airshipctl/+/790023 https://review.opendev.org/c/airship/airshipctl/+/790026 Thank you in advance! | 13:29 |
| airship-irc-bot | <cmti95035> Hi, there, I was trying the Steps to build a Local All-in-one VM Environment on an AWS instance. I started with a fresh created Ubuntu 18.04 instance and follow the steps in the dev guide page. It failed on the step when running the ./tools/gate/20_run_gate_runner.sh, specifically from ./tools/deployment/25_deploy_ephemeral_node.sh: | 19:02 |
| airship-irc-bot | <cmti95035> + echo '1: Retrying to reach the apiserver' 1: Retrying to reach the apiserver + sleep 60 + '[' 1 -ge 30 ] + timeout 20 kubectl --kubeconfig /kubeconfig --context ephemeral-cluster get node Unable to connect to the server: dial tcp 10.23.25.101:6443: connect: no route to host | 19:03 |
| airship-irc-bot | <kk6740> did it timeout? or it is still in the process? | 19:03 |
| airship-irc-bot | <cmti95035> It retried 30 times and failed the operation. Can anyone advise on what might go wrong here? | 19:04 |
| airship-irc-bot | <cmti95035> It times out | 19:04 |
| airship-irc-bot | <kk6740> can u run kvm-ok command on the host? | 19:04 |
| airship-irc-bot | <cmti95035> hold on. let me try | 19:06 |
| airship-irc-bot | <cmti95035> I have to run sudo /usr/sbin/kvm-ok and the output is INFO: Your CPU does not support KVM extensions KVM acceleration can NOT be used | 19:14 |
| airship-irc-bot | <kk6740> u need KVM to deploy a site | 19:14 |
| airship-irc-bot | <kk6740> so u need to make sure that ur VM supports kvm and nested virt | 19:15 |
| airship-irc-bot | <cmti95035> This is a bit confusing -- since Airship 2.0 supports deployment over public cloud, this would be an additional requirement to the VMs to be used. As it shows that "Your CPU does not support KVM extensions" which suggests certain instances won't support KVM, then how would I know upfront what to choose to enforce a given instance can support KVM? | 19:18 |
| airship-irc-bot | <kk6740> ah, this environment is designed to emulate baremetal deployment | 19:19 |
| airship-irc-bot | <cmti95035> BTW, is it mentioned anywhere in the documentation that a VM has to support KVM? | 19:19 |
| airship-irc-bot | <kk6740> so that is why KVM is required | 19:19 |
| airship-irc-bot | <cmti95035> oh I see... | 19:20 |
| airship-irc-bot | <kk6740> you certainly can craft an AWS site with airship 2.0. but we dont have it predefined at the moment. I think @sidney.shiba was working on Azure and GCP environments | 19:21 |
| airship-irc-bot | <kk6740> but all in one VM is basically a development environment that emulates real baremetal hosts via virtual machines | 19:22 |
| airship-irc-bot | <cmti95035> Good to know all this -- thanks a lot for your help on this! | 19:23 |
| airship-irc-bot | <sidney.shiba> @cmti95035, I have deployed all-in-one VM instantiated in Azure cloud. The requirement is that the VM you choose must support nested-virtualization. I have tested on Azure VM size Standard D4s v3 (4vcpus, 16 GB mem). This is deploying the baremetal operator k8s cluster. If you want to deploy CAPA (Cluster API AWS), meaning deploying controlplane and worker nodes as AWS VMs, it would need to add CAPA manifests to the | 19:31 |
| airship-irc-bot | "airshipctl", which should be fairly easy. Note, we have integrated Azure, GCP, Openstack and Docker with airship so far and of course metal3 baremetal. | 19:31 |
| airship-irc-bot | <cmti95035> @sidney.shiba Thanks for the explanation. Are you suggesting that by adding CAPA manifest to the airshipctl it will automatically take care of the VMs to support nested-virtualization? | 19:34 |
| airship-irc-bot | <cmti95035> Also just curious -- AWS is by far the largest public cloud provider. Seems to me you might want to support AWS upfront instead of Azure/GCP. | 19:35 |
| airship-irc-bot | <sidney.shiba> No. For all-in-one you manually instantiate a (support nested-virtualization) VM and you run the baremetal deployment scripts inside this VM. | 19:38 |
| airship-irc-bot | <cmti95035> Would someone point me to an "airshipctl plan" example? I would like to see how a life cycle management might look like. Thanks! | 19:48 |
| airship-irc-bot | <kk6740> here is example, where we add airshipctl plan to deploy test baremetal environment | 19:51 |
| airship-irc-bot | <kk6740> https://review.opendev.org/c/airship/airshipctl/+/791715/17/manifests/type/gating/plan.yaml | 19:51 |
| airship-irc-bot | <kk6740> here is the plan used to test other component called VINO https://opendev.org/airship/vino/src/branch/master/config/phases/plans.yaml | 19:52 |
| airship-irc-bot | <kk6740> there is working happening right now to introduce plan per site type that you want to deploy. For example it would be, `aws-plan` or `gcp-plan` | 19:53 |
| airship-irc-bot | <cmti95035> Cool! I'll take a look. Any plan regarding upgrade of a site as I consider it's an important piece of life cycle management | 19:54 |
| airship-irc-bot | <kk6740> we plan to have multiple plans per site, so for example one plan would be responsible for deployment, second plan for upgrading, and 3rd one would focus on workloads | 19:55 |
| airship-irc-bot | <mf4716> Treasuremap Core Reviews and WF request for Niharikas/Sreejith's PS in support of 2.1 - https://review.opendev.org/c/airship/treasuremap/+/791979 | 19:55 |
| airship-irc-bot | <kk6740> we have all the tools required to craft those plans, however the crafting itself is in the progress atm | 19:56 |
| airship-irc-bot | <cmti95035> So this aws-plan or gcp-plan is like a template predefined for the deployment phases? One just needs to fill in the credentials? I still see a "TODO: Azure subscription credentials to be passed as environment variables" in the "Airshipctl and Azure Cloud Platform Integration" page | 19:56 |
| airship-irc-bot | <kk6740> the idea is that you will only have to export ur crednetials as ENV variables , yes | 19:57 |
| airship-irc-bot | <kk6740> well, you would also be able to specify number of clusters u want, how many workers, which flavors etc, but that would be done decleratively as variables in your site manfiest | 19:58 |
| airship-irc-bot | <cmti95035> That makes sense | 20:01 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!