-
Notifications
You must be signed in to change notification settings - Fork 18.8k
Bump go 1.13.0 #39549
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
Bump go 1.13.0 #39549
Conversation
Dockerfile
Outdated
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
opened checkpoint-restore/criu#743 for this as well
Windows build is failing, but no details what exactly is failing 😞
|
And new
LOL, I don't see a difference; diff --git a/pkg/parsers/kernel/kernel_windows.go b/pkg/parsers/kernel/kernel_windows.go
index b7b15a1fd2..a0712ce633 100644
--- a/pkg/parsers/kernel/kernel_windows.go
+++ b/pkg/parsers/kernel/kernel_windows.go
@@ -44,7 +44,7 @@ func GetKernelVersion() (*VersionInfo, error) {
}
KVI.major = int(dwVersion & 0xFF)
- KVI.minor = int((dwVersion & 0XFF00) >> 8)
+ KVI.minor = int((dwVersion & 0xFF00) >> 8)
KVI.build = int((dwVersion & 0xFFFF0000) >> 16)
return KVI, nil Oh! lowercase |
Failures on Experimental, Janky, PowerPC, and s390x: https://jenkins.dockerproject.org/job/Docker-PRs/54911/console
Failures on Windows RS1 (fails to b 10BC0 uild) https://jenkins.dockerproject.org/job/Docker-PRs-WoW-RS1/25943/console
Failures on Windows RS5 (fails to build) https://jenkins.dockerproject.org/job/Docker-PRs-WoW-RS5-Process/3036/console
|
This one passes locally;
|
Logs for the failing And all logs of the tests and daemon: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
forcing IPv4 for the ping; see if that's making a difference
One thing I noticed on my machine (Docker for Mac) is that starting the daemon is really, really slow; looks like creating iptables rules takes a lot longer;
That's almost 7 seconds! If I start the daemon with
|
Switching to master, and using Go 1.12.7 on stretch;
And without iptables:
|
This branch again, using Go
And without iptables
So looks to be due to buster (perhaps the iptables version in buster switching to nftables?)
|
Added a bit of debugging for ``, and printing
Nothing that really stands out there |
Ok, found the problem (w.r.t. the failing tests); it's indeed because the CI machines run on Ubuntu 16.04, which uses legacy iptables, whereas the Debian buster image uses Added a hacky few lines to the https://wiki.debian.org/iptables
|
Looks like the performance issue is due to nftables; Switch to iptables-legacy:
Switch to iptables-nft:
|
switching to iptables-legacy fixes the tests;
|
hack/dind
Outdated
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Shouldn't this happen in the Dockerfile
instead so we don't accidentally apply this change somewhere less harmless than a container like on a contributor's host machine?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It's really a quick hack to see if it did the job
Surprisingly; when I did it as a RUN
step, it gave a "permission denied; need to be root
"
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oh that's really odd -- AFAICT, we're not using USER
or anything so it already should be root
. 😓
(And doing update-alternatives
shouldn't be running iptables
or anything that might need additional privileges.)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That's what I would expect as well; didn't look closely yet, but I suspect it's calling iptables (perhaps just to check if it's installed?)
Arg; forgot to remove
|
Instead of removing it, I'd recommend switching it to be |
Yes, that's what I actually meant 😅 - updated 🤞 |
Will have to look into the Windows build failure; the file that it marks missing is on the repository; https://github.com/moby/moby/blob/master/vendor/cloud.google.com/go/compute/metadata/metadata.go
|
73e405d
to
eb4289e
Compare
Yes, was actually thinking of separating it from this PR (although it is "accepted" by older versions of Go) |
8f9487c
to
ccdaf19
Compare
ccdaf19
to
c4e96d8
Compare
Interesting; this is why
Looking at this, it looks like
|
c4e96d8
to
810e9b2
Compare
ha! looks like Still wondering why it switched to use |
810e9b2
to
2e19350
Compare
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
2e19350
to
38e4ae3
Compare
@kolyshkin @tiborvass @tianon this is all green now; ready for review! |
@seemethere @zelahi I'm afraid if we merge this, release scripts somewhere will not add the GO111MODULE=off. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM 😄
Based on top of / depends on (when backporting, these are also needed):
I'll rebase once those are merged
testing if things work or break