A change of the active FedWiki site that never takes effect still starts the wait before the next change #328

Closed
opened 2026-11-10 08:25:58 +00:00 by cgalo5758 · 1 comment
Owner

What happens

An organization with more FedWiki sites than its plan keeps active can change which site is active: activating a parked site parks another one in its place. After a change, the organization waits a set period before it can change again (fedwiki-swap-cooldown, 30 days by default). The console starts that wait in the transaction that records the change, before the change's workflow runs, and nothing ever takes it back. A change that never takes effect starts the wait all the same:

  • When the change's workflow cannot be started, the member sees "Couldn't change the active site. Try again." A retry is refused with "The active site changes once every 30 day(s); try again in about 30 day(s)."
  • When the workflow runs and the change fails or is declined, both sites keep their old status, and the wait runs anyway.

Bringing back a site that a plan downgrade parked starts the same wait the same way.

What should happen

Only a change that took effect starts the wait. A change that never started, failed or was declined leaves the wait where it was, so a retry can follow "Try again."

Where

internal/integrations/fedwiki/web/site_actions.go: swapActiveSite and changeStatus both call UpsertSiteSwapPolicy in the recording transaction. fedwiki.site_swap_policy holds one time per organization, and the store has no query that clears or restores it.

Why it matters

A member who follows the console's own "Try again" is shut out of changing their active site for a month, for a change that never happened.

Related: #298, where a failed read of the same time is treated as no earlier change.

## What happens An organization with more FedWiki sites than its plan keeps active can change which site is active: activating a parked site parks another one in its place. After a change, the organization waits a set period before it can change again (`fedwiki-swap-cooldown`, 30 days by default). The console starts that wait in the transaction that records the change, before the change's workflow runs, and nothing ever takes it back. A change that never takes effect starts the wait all the same: - When the change's workflow cannot be started, the member sees "Couldn't change the active site. Try again." A retry is refused with "The active site changes once every 30 day(s); try again in about 30 day(s)." - When the workflow runs and the change fails or is declined, both sites keep their old status, and the wait runs anyway. Bringing back a site that a plan downgrade parked starts the same wait the same way. ## What should happen Only a change that took effect starts the wait. A change that never started, failed or was declined leaves the wait where it was, so a retry can follow "Try again." ## Where `internal/integrations/fedwiki/web/site_actions.go`: `swapActiveSite` and `changeStatus` both call `UpsertSiteSwapPolicy` in the recording transaction. `fedwiki.site_swap_policy` holds one time per organization, and the store has no query that clears or restores it. ## Why it matters A member who follows the console's own "Try again" is shut out of changing their active site for a month, for a change that never happened. Related: #298, where a failed read of the same time is treated as no earlier change.
cgalo5758 added the
kind
bug
area/fedwikiarea/member-ui
labels 2026-11-10 08:25:58 +00:00
Author
Owner

Fixed in 153b2812.

The wait between changes of the active site now starts only when a change succeeds. The step that records the success writes it, in the same transaction as the sites' new statuses. A change that cannot start, fails or is declined leaves the wait as it was, so a member can follow "Try again." at once.

Recording a change no longer starts the wait; it only reads it. While another change of the organization's active site is in progress, a change is refused with "A change to your active site is already in progress", whether or not a wait is configured.

Bringing back a site that a downgrade parked follows the same rule: it starts the wait only when it succeeds.

Fixed in 153b2812. The wait between changes of the active site now starts only when a change succeeds. The step that records the success writes it, in the same transaction as the sites' new statuses. A change that cannot start, fails or is declined leaves the wait as it was, so a member can follow "Try again." at once. Recording a change no longer starts the wait; it only reads it. While another change of the organization's active site is in progress, a change is refused with "A change to your active site is already in progress", whether or not a wait is configured. Bringing back a site that a downgrade parked follows the same rule: it starts the wait only when it succeeds.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: wiki-cafe/member-console#328