Relationships
#4878 update.restart-daemons restarts a pre-existing worker daemon with no stable cache dir into an enrollment loop
Opened by stack72 · 11/18/2026· Shipped 11/18/2026
A worker daemon enabled before swamp-club/swamp#3855 (swamp-club#4862, #4870) has no SWAMP_WORKER_CACHE_DIR in its systemd unit or launchd plist, so each start uses a fresh temp cache and a new machine id. With update.restart-daemons enabled, a background update restarts it, the single-enrollment token refuses the new machine id (enrollment allowance exhausted), and systemd/launchd keep retrying. The autoupdate log still records daemonsRestarted: [worker], so status reads as a success.
Expected: the worker's restart() leaves such a daemon running on the old binary and records a restart error telling the operator to re-run swamp worker daemon enable with a new token, so swamp update --setup-auto status shows it. A unit that sets SWAMP_WORKER_CACHE_DIR (the new default, or an explicit --cache-dir from any release) restarts as before.
Raised by the CI adversarial review on swamp-club/swamp#3855.
Shipped
Click a lifecycle step above to view its details.
Sign in to post a ripple.