-
Notifications
You must be signed in to change notification settings - Fork 38.6k
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
GC old dead containers and unused images #157
Comments
@roberthbailey will handle this one. |
FYI, Docker is also contemplating GC of unreachable images: That shouldn't stop us from doing something, but we should prioritize complementary functionality. |
Ping on this. It really makes things difficult when your stuff is crashlooping. |
This is slowly percolating to the top of my TODO list. ETA to start on it is before end of the week. |
I've written a small cli program that gets the output of ListContainers using the go docker remote api to see all of the dead and alive containers (this is the same output as 'docker ps -a'). Unfortunately, it looks like the only way to tell the status of a container is to use string parsing, as the status field is something like "Up 3 seconds" or "Exited (0) 53 minutes ago" rather than an actual type. This is going to be fragile. |
You should be able to look at the 'inspect' output for each container, no?
On Wed, Sep 3, 2014 at 3:16 PM, roberthbailey notifications@github.com
|
That sounds much better than what I was seeing. Let me give it a try... |
Is this being actively pursued? Should it be reassigned? |
Not currently being pursued. Feel free to reassign. |
This should be closed by #2022, yes? |
@ryaneleary That only addressed containers, not images. |
I am reopening this issue since I believe #3097 doesn't address image gc issue. See #3264 for detail reasons. In a short version, current support of gc of image depending on docker's "dangling=true", which gives us proper unused images on a node. Instead it only returns a list of image layers without tag. |
Created #5457 and assigned to myself. |
Integrated the imageManager into the Kubelet and applies the garbage collection policy every 5 minutes. The default policy allows up to 90% disk usage, after which images are garbage collected to bring limit back down to 80%. Fixes kubernetes#157.
Integrated the imageManager into the Kubelet and applies the garbage collection policy every 5 minutes. The default policy allows up to 90% disk usage, after which images are garbage collected to bring limit back down to 80%. Fixes #157.
add sig governance doc
Omit error if dir invalid
…stions Add brief section about external dependencies
…tment_for_test_framework No special treatment for test framework
…lop' [feature] support emptydir kubernetes#157 kubernetes#172 See merge request whale/kubernetes!27
As far as I can tell, we never call DELETE/RemoveContainer nor RemoveImage.
We should keep containers around for some amount of time for inspection, copying files out, etc., but should eventually clean them up.
We could treat images as a cache, evicting on LRU when we reach some predetermined space limit -- ideally before the disk just fills up.
The text was updated successfully, but these errors were encountered: