[podman-5.8] container storage cleanup work#1007
Conversation
We already have 3 users, and we will need another one. Should not change behavior. Signed-off-by: Miloslav Trmač <mitr@redhat.com> (cherry picked from commit 7cac518) Signed-off-by: Tom Sweeney <tsweeney@redhat.com>
Should not change behavior. Signed-off-by: Miloslav Trmač <mitr@redhat.com> (cherry picked from commit 4e5f9d5) Signed-off-by: Tom Sweeney <tsweeney@redhat.com>
- The graph driver contents are locked by layerStore's locking, not by the graph driver lock - Because we dropped the layer store lock, a layer might have legitimately been added in the meantime; don't treat is as unaccounted. Signed-off-by: Miloslav Trmač <mitr@redhat.com> (cherry picked from commit 8311e6c) Signed-off-by: Tom Sweeney <tsweeney@redhat.com>
…c/check-refactor Refactor `Store.Check` (cherry picked from commit c4edeb1) Signed-off-by: Tom Sweeney <tsweeney@redhat.com>
I failed to push the latest version of my code in podman-container-tools#720. Should not change behavior. Signed-off-by: Miloslav Trmač <mitr@redhat.com> (cherry picked from commit 0ebd8a2) Signed-off-by: Tom Sweeney <tsweeney@redhat.com>
As the title says. Signed-off-by: Tom Sweeney <tsweeney@redhat.com>
Luap99
left a comment
Luap99
left a comment
There was a problem hiding this comment.
Note we do not have the full CI running on this branch, as I did not bother backporting it here so far, I can look into it if needed.
But one thing that does run is the linter and that suggest you changes here do not compile. So please check that locally
$ go build ./storage/...
# go.podman.io/storage
storage/check.go:304:27: undefined: diffReader
But also I ado not follow the commits that have been picked, 6dc3cd5 cherry-picked the merge commit which I think is wrong, we want the original commits picked only. OR at least that is how we have been doing this in the past.
mtrmac
left a comment
mtrmac
left a comment
There was a problem hiding this comment.
From a literally 1-minute look, the locking fixes are probably relevant; the “Refactor Store.Check” branch/PR should be a no-op, so I think an unnecessary risk on a stable branch (if nothing else depends on that those changes).
|
(working on some proper CI here #1010, still WIP) |
3 participants
This backports a number of fixes relating to customer storage cleanup, especially noted in https://redhat.atlassian.net/browse/ACCELFIX-649
This cherry-picks from the commits from these PRs:
Store.Check#723Fixes: https://redhat.atlassian.net/browse/RHEL-173988, https://redhat.atlassian.net/browse/RHEL-173842
Once storage is vendored into Podman v5.8-rhel