Ctrl-Cing at the wrong moment a S3 transfer can SEGSIGV #1023
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
awaiting
author
awaiting
contributors
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#1023
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?
Describe the bug
Ctrl-Cing at the wrong moment a S3 transfer can lead to segmentation faults.
Steps To Reproduce
nix copyExpected behavior
No segfaults.
nix --versionoutputHEAD as of
91867941fa+cl/4514+cl/4513.Additional context
Here's the stacktrace, it seems plausible to me that we are shutting down some global facility that AWS uses but we are unable to "cancel" the work being done, so when things are continuing their life, they are exploding.
An minor issue.
since there is no way to cancel single async operations of an s3 client this may be unfixable in principle. the best workaround we can hope for may be to keep the s3 client alive while operations are running by shoving it into a
shared_ptrand capturing a copy of that into operation callbacks, but that makes it impossible to stop operations since the only avenue for this offered by the sdk is to pause the queue, keeping the client alive forever.tl;dr the aws sdk is ... very badly suited for our use case and we should use something else
well, fortunately we want to get rid of it. so i guess we can disregard this bug and consider it blocked on someone funding getting rid of the problem library.
@jade I have plans for this, so let's keep it open so I can "not" forget about it.