Skip to content

cmdlib: Drop unsafety for cache disk#3619

Merged
dustymabe merged 1 commit intocoreos:mainfrom
cgwalters:cache-drop-unsafe
Sep 14, 2023
Merged

cmdlib: Drop unsafety for cache disk#3619
dustymabe merged 1 commit intocoreos:mainfrom
cgwalters:cache-drop-unsafe

Conversation

@cgwalters
Copy link
Member

This can't really be safe here because we have to ensure that the stuff we write makes it to the filesystem image between a separated fetch + build process.

cc #3591 which I suspect this will fix, but didn't verify.

This can't really be safe here because we have to ensure
that the stuff we write makes it to the filesystem image
between a separated fetch + build process.

cc coreos#3591
which I suspect this will fix, but didn't verify.
Copy link
Member

@dustymabe dustymabe left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@dustymabe dustymabe enabled auto-merge (rebase) September 14, 2023 21:30
@dustymabe dustymabe merged commit 25a1ae4 into coreos:main Sep 14, 2023
dustymabe added a commit to dustymabe/fedora-coreos-pipeline that referenced this pull request Sep 16, 2023
Some recent changes like droping unsafety for the cache disk [1]
and also adding a few more root reprovision tests that get run
on the ppc64le platform [2] are causing us to bump up into the
2 hour time limit. Let's bump it.

[1] coreos/coreos-assembler#3619
[2] coreos/fedora-coreos-config#2597
jlebon pushed a commit to coreos/fedora-coreos-pipeline that referenced this pull request Sep 16, 2023
Some recent changes like droping unsafety for the cache disk [1]
and also adding a few more root reprovision tests that get run
on the ppc64le platform [2] are causing us to bump up into the
2 hour time limit. Let's bump it.

[1] coreos/coreos-assembler#3619
[2] coreos/fedora-coreos-config#2597
@jlebon
Copy link
Member

jlebon commented Sep 21, 2023

Seems like we're still hitting this: #3631 (comment)

Wonder if we should revert this given that (we think) it has slowed down the pipeline. Is it truly unsafe to use unsafe in an orderly shutdown?

@jlebon
Copy link
Member

jlebon commented Oct 25, 2023

I do think cache=unsafe is safe for our use case since we do shut down the supermin VM. But that said, I don't understand this PR; our pipelines don't use --transient, so wouldn't this patch have had no effect anyway on #3591?

@cgwalters
Copy link
Member Author

I do think cache=unsafe is safe for our use case since we do shut down the supermin VM.

Not sure we do that in an orderly way though reliably? Needs investigation.

Wonder if we should revert this given that (we think) it has slowed down the pipeline.

I have no opposition to that.

our pipelines don't use --transient,

Yeah, fair.

aaradhak pushed a commit to aaradhak/fedora-coreos-pipeline that referenced this pull request May 14, 2024
Some recent changes like droping unsafety for the cache disk [1]
and also adding a few more root reprovision tests that get run
on the ppc64le platform [2] are causing us to bump up into the
2 hour time limit. Let's bump it.

[1] coreos/coreos-assembler#3619
[2] coreos/fedora-coreos-config#2597
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.

3 participants