build: Harden flaky Aeron tests in CI #32242
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
patriknw
commented
Nov 27, 2023
| @@ -215,8 +215,6 @@ jobs: | |||
| -Dakka.test.tags.exclude=gh-exclude,gh-exclude-aeron,timing \ | |||
| -Dakka.test.multi-in-test=false \ | |||
| -Dakka.cluster.assert=on \ | |||
| -Daeron.dir=/opt/volumes/media-driver \ | |||
| -Daeron.term.buffer.length=33554432 \ | |||
| clean ${{ matrix.command }} | |||
There was a problem hiding this comment.
This job is not in Kubernetes. Might have same problem with too small /dev/shm. Let me try...
There was a problem hiding this comment.
Plenty of space, no problem.
Filesystem Size Used Avail Use% Mounted on
/dev/root 84G 62G 22G 74% /
tmpfs 7.9G 172K 7.9G 1% /dev/shm
tmpfs 3.2G 1.1M 3.2G 1% /run
tmpfs 5.0M 0 5.0M 0% /run/lock
/dev/sdb15 105M 6.1M 99M 6% /boot/efi
/dev/sda1 63G 4.1G 56G 7% /mnt
tmpfs 1.6G 12K 1.6G 1% /run/user/1001
octonato
reviewed
Nov 27, 2023
.github/workflows/multi-node.yml
Outdated
| @@ -147,7 +144,8 @@ jobs: | |||
| gcloud components install gke-gcloud-auth-plugin | |||
| gcloud config set compute/region us-central1 | |||
| gcloud config set compute/zone us-central1-c | |||
| ./kubernetes/create-cluster-gke.sh "akka-artery-aeron-cluster-${GITHUB_RUN_ID}" | |||
| gcloud container clusters get-credentials akka-artery-aeron-cluster-test --zone us-central1-c --project akka-team | |||
| # ./kubernetes/create-cluster-gke.sh "akka-artery-aeron-cluster-test" | |||
There was a problem hiding this comment.
is this intentional? Not calling the script to create the cluster?
There was a problem hiding this comment.
leftover from my testing, thanks
patriknw
force-pushed
the
wip-dev-shm-patriknw
branch
from
77334b7
to
a50c9e5
Compare
* increase /dev/shm and use that (by default) * use default term buffer size * increase cpu requests, shouldn't matter but corresponds to what we want to use, 2 pods per node
patriknw
force-pushed
the
wip-dev-shm-patriknw
branch
from
a50c9e5
to
bf1d4f0
Compare
* more memory request * separate Aeron run in another workflow to make such test failures more clear
|
There was an error: "insuffiient usable storage for new log of ". I have increased it. I don't know if it accumulates when running all tests? It's supposed to delete the files on shutdown. |
|
I separated the aeron run in separate workflow. I hope that shows up so I can trigger a manual run if I merge this? |
|
a successful run in https://github.com/akka/akka/actions/runs/7015560435 |
This was referenced Nov 28, 2023
He-Pin
pushed a commit
to He-Pin/akka
that referenced
this pull request
Jan 7, 2024
* increase /dev/shm and use that (by default) * use default term buffer size * increase cpu requests, shouldn't matter but corresponds to what we want to use, 2 pods per node * more memory request * separate Aeron run in another workflow to make such test failures more clear
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This looks very promising. I have tried in a gke cluster. Verified with
df -h. It was 64 MB and now 1G.No more "Scheduled sending of heartbeat was delayed".
This wasn't possible when we tried last time #30601