-
Notifications
You must be signed in to change notification settings - Fork 38.7k
ci: Switch remaining Linux tasks to self-hosted #21652
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
Conversation
12ddc12 to
2f3552f
Compare
|
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers. ReviewsSee the guideline for information on the review process.
If your review is incorrectly listed, please react with 👎 to this comment and the bot will ignore it on the next update. ConflictsReviewers, this pull request conflicts with the following ones:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first. |
9dd511a to
a497f98
Compare
fa339d8 to
fa28b13
Compare
411b023 to
848fbc9
Compare
beb0b0e to
b83e83c
Compare
dergoegge
left a comment
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.
ACK fa8e89d
|
concept ACK fa8e89d Questions: Is Win64 native still running on cirrus? The code changes look like it shouldnt be but it has a warning:
The MSan task took 2.5 hours! yow. Is the bottleneck there just hardware? |
Yes, sorry for the confusion. This is only about Linux. I've adjusted the title. There is another pull about the msvc on Windows. |
On the first run it will build llvm/clang/msan + depends + Bitcoin Core on a fresh cache. See also the comment in the cirrus yaml. |
hebasto
left a comment
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.
ACK fa8e89d.
Observing very low Ccache hit rate. For example, in https://cirrus-ci.com/task/6728143201894400:
ccache version 3.7.7
cache directory /tmp/ccache_dir
primary config /tmp/ccache_dir/ccache.conf
secondary config (readonly) /etc/ccache.conf
stats updated Fri Aug 18 17:06:37 2023
cache hit (direct) 110
cache hit (preprocessed) 15
cache miss 737
cache hit rate 14.50 %
called for link 12
cleanups performed 0
files in cache 1489
cache size 33.6 MB
max cache size 200.0 MB
|
I wonder why the hit rate is non-zero. On the first run, the cache is empty, obviously. Maybe a leftover ccache from a previous push to this pull request? |
Can you please re-run all Cirrus jobs to test caching facilities? |
There are multiple workers per label type and multiple tasks using the same label type. Thus, it will take many re-runs to populate the initial cache on all workers. |
|
I could limit the label type to each task, but that would be micro-managment. And free resources on one label type can not be used on another label type. |
Fair enough. We can observe caching quality in-progress later. |
Sounds like we need to increase |
Why? Each task has its own name (space). The CI is now doing exactly what happens when you run the CI locally. |
IIUC, multiple tasks share the same |
No, and as I said this is unrelated to the changes here. If there is a bug in the CI system when running locally, it should be fixed separately, not as part of this pull. See also: |
|
https://github.com/bitcoin/bitcoin/runs/16142913468 -- timeouts. Probably unrelated... |
|
yeah, the blob is missing. It would be good to switch the mirror URL to something that supports ipv6 and is maintained. Hint: #17704 |
|
Apart from the blob missing, this seems to be an odd issue, because it works outside of |
Cirrus CI will be capping the free compute soon. For now, switch more tasks to persistent worker, as recommended by Cirrus CI.
(See slightly related discussion in #28098)