[
https://issues.apache.org/jira/browse/FLINK-31210?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17694639#comment-17694639
]
Matthias Pohl edited comment on FLINK-31210 at 2/28/23 5:00 PM:
----------------------------------------------------------------
Ok, the overall pattern suggests that it's only affecting e2e_2 stages.
Additionally, it only seems to appear on the {{release-1.16}} branch.
It started appearing with {{5770182e}} quite consistently with
[20230223.2|https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=46434&view=results]
on Thu at 1:15 AM.
It's not happening on newer branches because we removed elasticsearch in 1.17.
I would have assumed it to happen on {{release-1.15}}, though, because we had
[20230224.2|https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=46490&view=results]
and
[20230223.17|https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=46477&view=results]
being triggered after Thu 1:15am
was (Author: mapohl):
Ok, the overall pattern suggests that it's only affecting e2e_2 stages.
Additionally, it only seems to appear on the {{release-1.16}} branch.
It started appearing with {{5770182e}} quite consistently with
[20230223.2|https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=46434&view=results]
on Thu at 1:15 AM.
It's not happening on newer branches because we removed elasticsearch in 1.17.
> Azure Pipelines report warning on "no space left on device" in Restore Docker
> images step
> -----------------------------------------------------------------------------------------
>
> Key: FLINK-31210
> URL: https://issues.apache.org/jira/browse/FLINK-31210
> Project: Flink
> Issue Type: Bug
> Components: Test Infrastructure
> Affects Versions: 1.16.1
> Reporter: Matthias Pohl
> Priority: Major
> Labels: test-stability
>
> We're experiencing "no space left on device" issues lately when restoring the
> Docker images. This doesn't make the builds fail, though. It's just reported
> as warning
> {code}
> Loaded image: docker.elastic.co/elasticsearch/elasticsearch:6.8.20
> ApplyLayer exit status 1 stdout: stderr: write
> /usr/share/elasticsearch/modules/ingest-geoip/GeoLite2-City.mmdb: no space
> left on device
> ##[error]Bash exited with code '1'.
> {code}
> All of this happens on Azure machines (not the Alibaba VMs). Therefore, I'm
> not sure whether there's much to do on our side.
> *
> [20230224.3|https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=46491&view=logs&j=f8e16326-dc75-5ba0-3e95-6178dd55bf6c&t=728e59c6-8078-53a8-7bbe-bb7b0b1f2c63]
> *
> [20230223.18|https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=46486&view=results]
> *
> [20230223.6|https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=46441&view=results]
> *
> [20230223.2|https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=46434&view=results]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
