Flaky tests: nix-copy and nix-copy-closure #490
Labels
No labels
Affects/CppNix
Affects/Nightly
Affects/Only nightly
Affects/Stable
Area/build-packaging
Area/cli
Area/evaluator
Area/fetching
Area/flakes
Area/language
Area/lix ci
Area/nix-eval-jobs
Area/profiles
Area/protocol
Area/releng
Area/remote-builds
Area/repl
Area/repl/debugger
Area/store
bug
Context
contributors
Context
drive-by
Context
maintainers
Context
RFD
crash 💥
Cross Compilation
devx
docs
Downstream Dependents
E/easy
E/hard
E/help wanted
E/reproducible
E/requires rearchitecture
Feature/S3
imported
Language/Bash
Language/C++
Language/NixLang
Language/Python
Language/Rust
Needs Langver
OS/Linux
OS/macOS
performance
regression
release-blocker
stability
Status
blocked
Status
invalid
Status
postponed
Status
wontfix
testing
testing/flakey
Topic/Large Scale Installations
ux
No milestone
No project
No assignees
3 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
lix-project/lix#490
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
The
nix-copyandnix-copy-closureNixOS tests intests/nixosseem to be flaky.Build failures here: https://buildbot.lix.systems/#/buildrequests/167235, https://buildbot.lix.systems/#/buildrequests/167245, https://buildbot.lix.systems/#/buildrequests/167036
nix-copy-closure = runNixOSTestFor "x86_64-linux" ./nix-copy-closure.nix;nix-copy = runNixOSTestFor "x86_64-linux" ./nix-copy.nix;lix-project/lix@0dc486a5bf/tests/nixos/nix-copy-closure.nixlix-project/lix@0dc486a5bf/tests/nixos/nix-copy.nixThese regularly break because of garbage collection, seemingly. I'm not sure what the exact reason is but I have my suspicions it could be either a gc bug or a design flaw.
i believe that the place where the paths are vanishing too early is on the coordinator.
we've taken a good long look and this problem seems to be straight up unfixable in the current test infra without removing bits of the test. the test requires that
nix copycan correctly pass ssh password prompts to the user (or at least it wants to require that, it does even less because the previous attempt was more broken), but because there's no sane way to wrap ssh in ways that simulate user interaction this is extremely busted.perhaps we should just deprecate password authentication for remote store interaction entirely. doing it through the daemon does not work anyway, and the mere chance that two ssh sessions may be open simulatenously and asking for passwords is a complete stopper too