Bump nix binary to 2.32.0 without caching #2563

OriolMunoz-da · 2025-10-06T13:58:23Z

Fixes https://github.com/DACH-NY/cn-test-failures/issues/5021 (hopefully, obviously)

Reverse of https://github.com/hyperledger-labs/splice/pull/1727/files:

All the fixes done there are untouched (so still there)
Newer nix version
Bumping cache versions again (assume 6 was broken, so 7) and add $HOME and the nix binary version to the cache key

Unfortunately keeping our caches (binary cache is fine) is proving to be very hard for me. Without them the timing impact is about 2m on average.

Latest version is 2.32
See this change done in 2.14 that is not documented but likely breaks part of our setup.

OriolMunoz-da · 2025-10-07T13:22:48Z

OriolMunoz-da · 2025-10-07T15:20:31Z

apps/sv/frontend/src/__tests__/governance/forms/set-amulet-rules-form.test.tsx

+      expect(screen.getByText(/Submission failed/)).toBeDefined();
+      expect(screen.getByText(/Service Unavailable/)).toBeDefined();
+    },
+    { timeout: 10000 }


bumping a timeout because it flaked

Signed-off-by: Oriol Muñoz <[email protected]>

…ight?" This reverts commit 4984cef. Signed-off-by: Oriol Muñoz <[email protected]>

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da · 2025-10-08T14:35:45Z

checkpoint: 6814e44 seems to work with a few cache-related things disabled
it's up to 2m slower than current main... let's see how far I get otherwise

Signed-off-by: Oriol Muñoz <[email protected]>

nicu-da · 2025-10-09T19:15:05Z

latest checkpoint: daf9b99 comparing times between here (there was a flake, but doesn't look related to nix) and some main run The avg time impact is of about 2s

but the setup is more than a minute in diff, right?

the diff is smaller because with the caches for some reason nix had a 10s time period where it did nothing, never managed to figure out why, but that's dependent on how many run with nix actions you have in a workflow. for the integration tests we usually have a lot less so the diff is a bit higher

OriolMunoz-da · 2025-10-10T10:01:40Z

@nicu-da
working-late typo:

The avg time impact is of about 2s

was supposed to be 2m... 🤦
that includes everything done in a test job, not just the nix setup

The unfortunate thing is that most test runs don't show a separate step for setting up nix, but rather it's included in the "Run tests" step, so I cannot check for individual times from gh

This reverts commit 8920f86. Signed-off-by: Oriol Muñoz <[email protected]>

This reverts commit ec92b34. Signed-off-by: Oriol Muñoz <[email protected]>

This reverts commit 8f58bad. Signed-off-by: Oriol Muñoz <[email protected]>

This reverts commit 6fc6702. Signed-off-by: Oriol Muñoz <[email protected]>

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da · 2025-10-10T10:16:58Z

.github/workflows/build.ts_cli_tests.yml


      - name: Check out repository code
        uses: actions/checkout@v4
-        if: inputs.commit_sha != ''


this and below are the only ones that had this... what for?

@isegall-da is that leftover? I could see it only being set when you run from a fork but then it should be used for every checkout and it's definitely not.

ping @isegall-da , don't want to merge without knowing this is 100% fine

Discussed with @OriolMunoz-da now, just copying the conclusion here as well:
In these tests we do:
Checkout the PR head commit to get the commit message first
check the static label
Checkout commit_sha

In tests that run split_tests first, we check the static label in split_tests which is why we don't have these separate steps there, and we just checkout commit_sha there.

(in ts_cli we flipped the order between (2) and (3) which is wrong, we'll fix that)

OriolMunoz-da · 2025-10-10T10:17:12Z

.github/workflows/build.ui_tests.yml


      - name: Check out repository code
        uses: actions/checkout@v4
-        if: inputs.commit_sha != ''


[ci] Signed-off-by: Oriol Muñoz <[email protected]>

Signed-off-by: Oriol Muñoz <[email protected]>

moritzkiefer-da

Nice thanks! Please hold off on merging until Martin and Nicu weigh in.

moritzkiefer-da · 2025-10-10T13:55:49Z

.github/workflows/build.ts_cli_tests.yml


      - name: Check out repository code
        uses: actions/checkout@v4
-        if: inputs.commit_sha != ''


@isegall-da is that leftover? I could see it only being set when you run from a fork but then it should be used for every checkout and it's definitely not.

moritzkiefer-da · 2025-10-10T13:56:58Z

.github/actions/nix/setup_nix/action.yml

          rsync -avi "/cache/nix/$cache_key/nix" $HOME/.config/
-          rsync -avi "/cache/nix/$cache_key/nix_store/var/" /nix/var
-          sudo mount --bind /cache/nix/$cache_key/nix_store/store /nix/store
+          # rsync -avi "/cache/nix/$cache_key/nix_store/var/" /nix/var


If we do uncomment that let's create an issue for it documenting how it fails and what you tried.

@martinflorian-da @nicu-da wdyt? I'm leaning towards saying getting an upgrade in with the slowdown and then moving the fix for this to the top of our prio list is better than to keep pushing out that upgrade for another few months (the pressure from slow CI also felt more so chances are higher we'll actually tackle it).

(With my limited context:) Agreed to merging this without this caching part, a bit worried about "just moving the fix to the top of our prio list". Can we give assign it to a concrete person already that will realistically get to it soon?

Agree with @martinflorian-da those 2 minutes were really annoying in the past, though I guess with the current CI timing it might be less so.

#2663 adding as TODOs

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da · 2025-10-10T15:36:53Z

Please hold off on merging until ...

...Monday, because nobody is dealing with any potential downfall today 😄

isegall-da · 2025-10-13T13:16:22Z

@OriolMunoz-da

I cannot check for individual times from gh

Note that GH has a "show timestamps" when you look at its output.

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da changed the title ~~Bump nix binary to 2.31.2~~ [wip] Bump nix binary to 2.31.2 Oct 6, 2025

OriolMunoz-da changed the title ~~[wip] Bump nix binary to 2.31.2~~ [wip] Bump nix binary to $SOMETHING Oct 7, 2025

OriolMunoz-da commented Oct 7, 2025

View reviewed changes

OriolMunoz-da added 23 commits October 8, 2025 13:04

bump nix and all cache versions

d013e04

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] run

607ecbd

Signed-off-by: Oriol Muñoz <[email protected]>

use a ridiculous cache version

a3f0a03

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] run

5f160a3

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] increase timeout on a test i just stumbled upon

d423b6b

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] so surely if I don't restore the cache, this will work right?

925e123

Signed-off-by: Oriol Muñoz <[email protected]>

Revert "[ci] so surely if I don't restore the cache, this will work r…

20fddda

…ight?" This reverts commit 4984cef. Signed-off-by: Oriol Muñoz <[email protected]>

[ci] try with 2.28.5 (because breaking changes might be causing this)

eb97931

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] ok why are you different my dude

f299366

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] there's no sane reason why you're different

6728a67

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] is this inherent to bumping nix or a version-related issue?

4b56865

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] ok now with 14

6aafc5c

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] attempt without GC

a02c1bb

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] try without mounting the damn thing

dcbf412

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] when did some verbose flags ever hurt anyone

bbecb2d

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] log version in run

60248a2

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] ok what happens if i create it beforehand

0f94215

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] probably won't work but moritz says i can try

10e8b5c

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] soilly me

f81368a

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] orint env

b74683b

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] add home to the cache key

bcb3d44

Signed-off-by: Oriol Muñoz <[email protected]>

Merge branch 'main' into oriol/nix-electric-boogaloo

8b83e26

[ci] bump to 2.32.0 (released yesterday) and add it to the cache key

6814e44

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da force-pushed the oriol/nix-electric-boogaloo branch from cda8f20 to 6814e44 Compare October 8, 2025 13:10

fix comment

a8aac68

Signed-off-by: Oriol Muñoz <[email protected]>

[ci] gh was broken so i can't trust that

8920f86

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da added 5 commits October 10, 2025 10:12

Revert "[ci] gh was broken so i can't trust that"

fd6a63a

This reverts commit 8920f86. Signed-off-by: Oriol Muñoz <[email protected]>

Revert "[ci] stupid test: is the mount the problem?"

e75ba96

This reverts commit ec92b34. Signed-off-by: Oriol Muñoz <[email protected]>

Revert "[ci] maybe exclude the DB too? (will this cache anything then?)"

faeb743

This reverts commit 8f58bad. Signed-off-by: Oriol Muñoz <[email protected]>

Revert "[ci] attempt with caches excluding personal profile stuff"

285dc2e

This reverts commit 6fc6702. Signed-off-by: Oriol Muñoz <[email protected]>

[ci] sensible cache_Version

a5115a4

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da changed the title ~~[wip] Bump nix binary to $SOMETHING~~ [wip] Bump nix binary to 2.32.0 Oct 10, 2025

OriolMunoz-da changed the title ~~[wip] Bump nix binary to 2.32.0~~ Bump nix binary to 2.32.0 Oct 10, 2025

OriolMunoz-da commented Oct 10, 2025

View reviewed changes

OriolMunoz-da added 2 commits October 10, 2025 10:44

Merge branch 'main' into oriol/nix-electric-boogaloo

39c6b7e

[ci] Signed-off-by: Oriol Muñoz <[email protected]>

[ci] fix error...

180665b

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da marked this pull request as ready for review October 10, 2025 13:34

OriolMunoz-da requested review from isegall-da and moritzkiefer-da October 10, 2025 13:34

moritzkiefer-da approved these changes Oct 10, 2025

View reviewed changes

OriolMunoz-da requested review from martinflorian-da and nicu-da October 10, 2025 14:01

OriolMunoz-da mentioned this pull request Oct 10, 2025

Restore caching in Nix #2663

Open

Add todos and [ci] for good measure

f39ce61

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da changed the title ~~Bump nix binary to 2.32.0~~ Bump nix binary to 2.32.0 without caching Oct 10, 2025

OriolMunoz-da added 2 commits October 13, 2025 15:09

Merge branch 'main' into oriol/nix-electric-boogaloo

a500bc5

[ci] itai review fixes

c9decd8

Signed-off-by: Oriol Muñoz <[email protected]>

OriolMunoz-da merged commit f333d10 into main Oct 14, 2025
57 checks passed

OriolMunoz-da deleted the oriol/nix-electric-boogaloo branch October 14, 2025 08:52

OriolMunoz-da mentioned this pull request Oct 14, 2025

Bump version of nix #1577

Open

Bump nix binary to 2.32.0 without caching #2563

Bump nix binary to 2.32.0 without caching #2563

Conversation

OriolMunoz-da commented Oct 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

OriolMunoz-da commented Oct 7, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

OriolMunoz-da commented Oct 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nicu-da commented Oct 9, 2025

Uh oh!

OriolMunoz-da commented Oct 10, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

moritzkiefer-da left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

OriolMunoz-da commented Oct 10, 2025

Uh oh!

isegall-da commented Oct 13, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

OriolMunoz-da commented Oct 6, 2025 •

edited

Loading

OriolMunoz-da commented Oct 8, 2025 •

edited

Loading