ci: dbg ccache #34477

1. 
   maflcko commented at 11:09 am on February 2, 2026: member

   see if 100% hit rate can be achieved
2. DrahtBot added the label Tests on Feb 2, 2026
3. maflcko marked this as a draft on Feb 2, 2026
4. 
   DrahtBot commented at 11:09 am on February 2, 2026: contributor

   The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

   Reviews

   See the guideline for information on the review process. A summary of reviews will appear here.

5. ci: dbg ccache e2038c6fd1
6. maflcko force-pushed on Feb 2, 2026
7. DrahtBot added the label CI failed on Feb 2, 2026
8. 
   DrahtBot commented at 11:25 am on February 2, 2026: contributor

   🚧 At least one of the CI tasks failed. Task lint: https://github.com/bitcoin/bitcoin/actions/runs/21587712045/job/62200305011 LLM reason (✨ experimental): Lint step failed due to a shellcheck SC2086 quoting warning in the lint run, causing the docker lint tool to exit with error.

   Try to run the tests locally, according to the documentation. However, a CI failure may still happen due to a number of reasons, for example:

   • Possibly due to a silent merge conflict (the changes in this pull request being incompatible with the current code in the target branch). If so, make sure to rebase on the latest commit of the target branch.

   • A sanitizer issue, which can only be found by compiling with the sanitizer and running the affected test.

   • An intermittent issue.

   Leave a comment here, if you need help tracking down a confusing failure.

9. ci: bump CCACHE_MAXSIZE to 2G

   With 15 runners we get 150GB of cache space to use, and we currently
   have 18 jobs using ccache.
   
   Although each run only generates ~ 200-300 MB of cache data on each run,
   the small the small size may be contributing to poor hitrate.
   
   Bump ccache to 2GB per job ~ --> 36GB of the total 150GB cache space to
   try and reduce any thrashaing and generally increase hitrate.
   
   (cherry picked from commit 580e9eefe39f2f94695e1a22437c6ac47a1664e6)

   d13c0bc03c
10. 
    willcl-ark commented at 12:29 pm on February 2, 2026: member

    In this run we see previous releases restore from

    0Cache restored from key: ccache-ci_native_previous_releases-21586991578
    

    Which is run https://github.com/bitcoin/bitcoin/actions/runs/21586991578 … from which no code should have changed.

    This saved the cache at https://github.com/bitcoin/bitcoin/actions/runs/21586991578/job/62197978169#step:12:109 and it’s unclear to me therefore why we see any cache misses at all.

    Cache should be sized at ~ 250MB so should fit within the 500MB cache size. The rationale behind me opening #34474 was that, without a larger cache perhaps we are seeing some kind of cache thrashing where:

    • early cache misses cause LRU evictions of would-be future cache hits when the cache was a bit too small.

    Giving room for more than two full compilations may help to fix any such thrashing. But I admit, I don’t have any concrete evidence for it so mainly going on a hunch of what might help to fix it.

    I will push a branch to try and get debug output from ccache around why it’s missing

11. 
    maflcko commented at 1:26 pm on February 2, 2026: member

    looks like the bump to 2GB fixed it, but yeah, it seems a bit odd that locally less is needed
12. maflcko closed this on Feb 2, 2026

    ---

13. maflcko deleted the branch on Feb 2, 2026
14. 
    willcl-ark commented at 2:06 pm on February 2, 2026: member

    looks like the bump to 2GB fixed it, but yeah, it seems a bit odd that locally less is needed

    Where are you seeing that? This run https://github.com/bitcoin/bitcoin/actions/runs/21585565985 does not look good to me at all?

15. 
    maflcko commented at 2:11 pm on February 2, 2026: member

    I think you are looking at the worng pull pull request. The result of this pull is https://github.com/bitcoin/bitcoin/actions/runs/21590004916?pr=34477, which shows all 100% after 2G.

    The exact same pull, without the 2G, shows a mixed result: https://github.com/bitcoin/bitcoin/actions/runs/21588179370, no?

Contributors
maflcko DrahtBot willcl-ark

Labels
Tests CI failed