Bug 1725524
| Summary: | m4.xlarge is not supported in ap-northeast-2b | ||
|---|---|---|---|
| Product: | OpenShift Container Platform | Reporter: | sheng.lao <shlao> |
| Component: | Installer | Assignee: | W. Trevor King <wking> |
| Installer sub component: | openshift-installer | QA Contact: | sheng.lao <shlao> |
| Status: | CLOSED ERRATA | Docs Contact: | |
| Severity: | high | ||
| Priority: | high | CC: | eparis, gpei, jialiu, sponnaga, wking, wmeng |
| Version: | 4.1.z | Keywords: | Regression, Reopened |
| Target Milestone: | --- | ||
| Target Release: | 4.1.z | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2019-10-16 18:07:59 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | 1725526 | ||
| Bug Blocks: | |||
|
Description
sheng.lao
2019-07-01 03:28:17 UTC
looks a lot like https://github.com/openshift/installer/pull/1786 but in a different region... ``` aws --region ap-northeast-2 ec2 describe-reserved-instances-offerings --instance-tenancy default --instance-type m4.xlarge --product-description 'Linux/UNIX' --filters Name=scope,Values='Availability Zone' | jq -r '[.ReservedInstancesOfferings[].AvailabilityZone] | sort | unique[]' ap-northeast-2a ap-northeast-2b ap-northeast-2c ``` It looks like locally AWS is reporting that the `m4.xlarge` instances are allowed for all AZs in ap-norteast-2 https://docs.aws.amazon.com/autoscaling/ec2/userguide/ts-as-instancelaunchfailure.html#ts-as-instancelaunchfailure-6 In this page, it says: Cause: The instance type associated with your launch configuration might not be currently available in the Availability Zones specified in your Auto Scaling group. Solution: Update your Auto Scaling group with the recommendations in the error message. I think that we should check the configuration created by installer of Auto Scaling group. Abhinav, you seemed to have looked at `ReservedInstancesOfferings` not On Demand instances. My assumption is that you can get a reserved instance in 2b, but you can't get On Demand. And we use On Demand. I obviously have no idea if this is something AWS will correct, or if we need to switch to m5 in that region, or what.... This bug was opened for tracking 4.1.z stream, QE opened a separate bug for 4.2 - 1725526, so move this bug to 4.1.z Per QE's testing with 4.1.0-0.nightly-2019-07-10-210957, this bug is still reproduced. https://github.com/openshift/installer/pull/2108 is backporting a prereq before we can backport https://github.com/openshift/installer/pull/1935 to the release-4.1 branch. I've filed bug 1734136 with "do we want to support ap-east-1 in 4.1.z?". That will determine whether we want to manually backport 1935 to release-4.1 (if we decide not to support ap-east-1) or whether we want to wait until pull 2108 lands to get a clean 1935 backport (if we decide to support ap-east-1). It looks like we are not going to fix this issue for 4.1.z Per our official document [1], 'ap-northeast-2' is supported. If this region does NOT supported, need remove the region from installer, and update our official document. Am I right? [1]: https://docs.openshift.com/container-platform/4.1/installing/installing_aws/installing-aws-account.html#installation-aws-regions_installing-aws-account I've filed https://github.com/openshift/installer/pull/2213 with the dirty backport, because the conclusion of bug 1734136 was "we don't want the ap-east-1 stuff we'd need for a clean backport in 4.1.z". (In reply to Johnny Liu from comment #18) > Per our official document [1], 'ap-northeast-2' is supported. > > If this region does NOT supported, need remove the region from installer, > and update our official document. Am I right? > > > [1]: > https://docs.openshift.com/container-platform/4.1/installing/installing_aws/ > installing-aws-account.html#installation-aws-regions_installing-aws-account At some point the users will be asked to use 4.2 instead of 4.1.. the z streams for 4.1 is for critical patches. And there's already a workaround even for 4.1 ie. set the instance type manually in the install-config.yaml. Or even use the default instance types but restrict to one zone. The region is supported, just the defaults are no longer correct. (In reply to Abhinav Dahiya from comment #20) > (In reply to Johnny Liu from comment #18) > At some point the users will be asked to use 4.2 instead of 4.1.. the z > streams for 4.1 is for critical patches. And there's already a workaround > even for 4.1 ie. set the instance type manually in the install-config.yaml. > Or even use the default instance types but restrict to one zone. > > The region is supported, just the defaults are no longer correct. If `openshift create cluster` workflow offers use of ap-northeast-2 and it's not possible to walk through the installation process we should either remove that region or address the machine type problems. We should not leave 4.1 in a state where the defaults fail. Verified with version: 4.1.0-0.nightly-2019-08-22-165647 Steps: # openshift-install create cluster --dir test1 ? Platform [Use arrows to move, type to filter] > aws ? Platform aws ? Region [Use arrows to move, type to filter, ? for more help] ap-northeast-1 (Tokyo) > ap-northeast-2 (Seoul) ... ... # sh check_cluster_health.sh Passed Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:3004 |