Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Run imagecleaner more aggressively #3229

Merged
merged 2 commits into from
Feb 18, 2025
Merged

Conversation

yuvipanda
Copy link
Contributor

The reason we had to revert #3228 was that the image cache was growing big, and so was the buildkit cache. So this meant images were being stored in three places:

  1. the buildkit cache (dind)
  2. the image cache (dind)
  3. the containerd store (by kubernetes, when it pulls)

1 and 2 were both set to about 300, so this caused the disk to fill up.

Here, I more aggressively prune the image cache, as it is simply not used when buildkit is used. Once I verify this works ok, I'll turn on buildkit building again and we see how it goes.

yuvipanda and others added 2 commits February 18, 2025 13:25
The reason we had to revert jupyterhub#3228
was that the image cache was growing big, and so was the buildkit
cache. So this meant images were being stored in *three* places:

1. the buildkit cache (dind)
2. the image cache (dind)
3. the containerd store (by kubernetes, when it pulls)

1 and 2 were both set to about 300, so this caused the disk to
fill up.

Here, I more aggressively prune the image cache, as it is simply
not used when buildkit is used. Once I verify this works ok, I'll
turn on buildkit building again and we see how it goes.
@yuvipanda yuvipanda changed the title hetzner: Run imagecleaner more aggressively Run imagecleaner more aggressively Feb 18, 2025
@yuvipanda yuvipanda merged commit 049ea77 into jupyterhub:main Feb 18, 2025
8 checks passed
yuvipanda added a commit to yuvipanda/mybinder.org-deploy that referenced this pull request Feb 19, 2025
When we aren't using disk on the host for anything else, kubernetes
can simply use as much of the disk as it wants for caching images.
The default for this is upto 85%. However, since we *do* use the disk
for our dind, we want to tune this to take up only half that size.

I've already applied this to the 2i2c-hetzner node

Follow-up to jupyterhub#3229
yuvipanda added a commit to yuvipanda/mybinder.org-deploy that referenced this pull request Feb 19, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant