KillMode=process
may leave subdaemons which hold locks to paths a user want to build later #563
Labels
No labels
Area/build-packaging
Area/cli
Area/evaluator
Area/fetching
Area/flakes
Area/language
Area/profiles
Area/protocol
Area/releng
Area/remote-builds
Area/repl
Area/store
bug
crash 💥
Cross Compilation
devx
docs
Downstream Dependents
E/easy
E/hard
E/help wanted
E/reproducible
E/requires rearchitecture
imported
Needs Langver
OS/Linux
OS/macOS
performance
regression
release-blocker
RFD
stability
Status
blocked
Status
invalid
Status
postponed
Status
wontfix
testing
testing/flakey
ux
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: lix-project/lix#563
Loading…
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
Consider a bug in Lix that would cause a subdaemon to deadlock somewhere, e.g. while downloading stuff, building paths, whatever.
Assume that subdaemon locks path A via flock.
Build a new Lix, install it on your system, restart nix-daemon. Observe that subdaemon didn't go away.
Retry what you wanted to have, i.e. rebuild path A, same bug but the new subdaemon is hanging on the flock call because the previous subdaemon still hold the path A flock.
Steps To Reproduce
KillMode=process
will leave the subdaemons around)How to recognize this?
nix-daemon $CLIENT_PID
in the process list)sudo strace -fp $DAEMON_PID
flock
syscalls, you may be running in an instance of this issue.Expected behavior
Rebuilding path A should work.
Additional context
Fixing this would probably require moving to
KillMode=mixed
and removing progressively the existence of subdaemons à la inetd.