[jira] [Resolved] (GEODE-7641) Windows instance creation fails in CI

["Robert Houghton (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/GEODE-7641?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Robert Houghton resolved GEODE-7641.
------------------------------------
    Fix Version/s: 1.12.0
       Resolution: Fixed

> Windows instance creation fails in CI
> -------------------------------------
>
>                 Key: GEODE-7641
>                 URL: https://issues.apache.org/jira/browse/GEODE-7641
>             Project: Geode
>          Issue Type: Bug
>          Components: ci
>            Reporter: Alexander Murmann
>            Assignee: Robert Houghton
>            Priority: Major
>             Fix For: 1.12.0
>
>          Time Spent: 50m
>  Remaining Estimate: 0h
>
> Windows job fail with some regularity because they fail to create the instance correctly. The pattern seems to be that the job retries for a prolonged time. After that it starts failing because the resource already exists. (Are we missing when the instance finally gets created?)
> Logs form timed out creation attempts:
> {code:java}
> 11:12:49Updated [https://www.googleapis.com/compute/beta/projects/apachegeode-ci/zones/us-central1-f/instances/heavy-lifter-fbdf2e36-c559-5e39-8bb8-9b298c208d83].
> 11:12:50ERROR: (gcloud.beta.compute.reset-windows-password) 
> 11:12:50The instance may not be ready for use. This can occur if the instance was
> 11:12:50recently created or if the instance is not running Windows.
> 11:12:50Please wait a few minutes and try again.
> 11:12:55.
> 11:12:55This command creates an account and sets an initial password for the
> 11:12:55user [geode] if the account does not already exist.
> 11:12:55If the account already exists, resetting the password can cause the
> 11:12:55LOSS OF ENCRYPTED DATA secured with the current password, including
> 11:12:55files and stored passwords.
> 11:12:55
> 11:12:55For more information, see:
> 11:12:55https://cloud.google.com/compute/docs/operating-systems/windows#reset
> 11:12:55
> 11:12:55Would you like to set or reset the password for [geode] (Y/n)?  
> 11:12:55Resetting and retrieving password for [geode] on [heavy-lifter-fbdf2e36-c559-5e39-8bb8-9b298c208d83]
> {code}
> The above repeats till the task times out.
>  Then it transitions to tasks that fails like shown below:
> {code:java}
> 11:13:09/tmp/build/5ca5b479/geode /tmp/build/5ca5b479
> 11:13:09/tmp/build/5ca5b479
> 11:13:09Hashed apache-develop-main-WindowsIntegrationTestOpenJDK11-build8-test11-job#1110 (74 chars) -> heavy-lifter-fbdf2e36-c559-5e39-8bb8-9b298c208d83 (49 chars)
> 11:13:10Deploying to zone us-central1-a
> 11:13:11Applying the following labels to the instance: instance_type=heavy-lifter,time-to-live=1576912391,job-name=windowsintegrationtestopenjdk11,pipeline-name=apache-develop-main,build-name=1110,sha=7ac2c4d24a2ccc45c0ba2b09754f9caeaaa4080f
> 11:13:13ERROR: (gcloud.compute.instances.create) Could not fetch resource:
> 11:13:13 - The resource 'projects/apachegeode-ci/zones/us-central1-f/instances/heavy-lifter-fbdf2e36-c559-5e39-8bb8-9b298c208d83' already exists
> {code}
> This seems to have been happening since quite a while. See [this run|https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-main/jobs/WindowsIntegrationTestOpenJDK11/builds/1110] for an example.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)