The file came in named 20260502000000, which sorts ahead of the TOTP attempt
limit from August. Knex records migrations in the order it runs them, so the
last row in the table is not the same on every instance: a fresh install runs
both in name order and ends on the August one, while an instance that already
had it runs only this one and ends here.
Pro reads that last row to recognise which Community version it is upgrading
from, and it can only carry one name per version. Renaming to a timestamp
after the August migration makes the last row the same everywhere.
The rename is safe while this is only a few commits old on master and in no
tag. Anyone who pulled in that window and already ran it would see knex offer
it a second time.
The validation added with the admin user checks reached further than intended.
The seed runs on every container start, not only on the first, and it threw on
three conditions that are normal on a second start.
The worst was the check for an existing user: after the first start the
default admin exists by definition, so every restart after that ended in a
container that would not come up. The insert is wrapped in a try and falls
back to an update precisely because it is meant to run again, and that path
was never reached.
The other two were narrower but the same shape. `validator.isEmail` rejects
`admin@localhost`, which is a common value in self-hosted setups, so an
instance that had been running for a year would stop coming up after the
upgrade.
None of it throws now. Bad input is reported on stderr and skipped, the way
an invalid username already was, and a username that belongs to someone else
is left alone rather than taken. `db:create-admin-user` keeps the strict
validation, which is where a person is there to read it and type again.
Reported against 2.2.1 by someone reading the source. Every finding held.
**Sign-in had no ceiling.** Failures were logged with the caller's address
and nothing more. Two counters now — one per address, one per account —
because the two attacks look different: one source working through many
accounts is caught by the first, many sources working on one account by the
second, and behind a proxy only the second still means anything.
The count is kept in the process that serves the request. PLANKA needs no
Redis and the stock deployment is one container; run several and each keeps
its own count, which multiplies the ceiling by their number. That trade is
written where the limits are configured.
**The second factor could be guessed at leisure.** Six digits, and a
pending token that stayed valid for its full ten minutes however many codes
were wrong. Wrong codes are now counted on the session row — in the
database, so the count survives a restart and holds across every process —
and when the budget is spent the session is destroyed. After that even the
right code is refused and the login starts over from the password.
**Avatars, background images and favicons** checked the token's signature
and nothing else, so a revoked session, a deactivated account or a changed
password all kept working there for as long as the signature lasted, which
is a year by default. The five checks the API makes now live in one helper
that both use, rather than the shortened copy that had drifted from it.
**A link attachment's favicon** was fetched from wherever the URL pointed.
Storing a link is harmless — it is a string the user typed — but fetching
its icon is a request the server makes to an address the user chose, and
whether an icon came back reported on what is reachable from inside the
network. Server-side fetches now refuse private, loopback and link-local
addresses, `169.254.169.254` among them. The attachment is still created:
linking to an internal wiki is a legitimate thing to do, and it was the
server's own request that had to stop.
**The signing key.** Our own compose file ships `notsecretkey`, and it is
printed in the documentation — so on any instance that copied it, anyone can
sign a token for any account. PLANKA now says so on every start, and keeps
saying it, along with a key that is missing or shorter than 32 characters.
The placeholder carries the warning inline, where it is copied from.
**The backup script** wrote password hashes, live sessions, TOTP secrets and
SMTP credentials to an unencrypted archive. `BACKUP_PASSPHRASE` now encrypts
it, and without one the script says what it just put on disk. It also says
what it is — an example for the stock compose stack, not a backup concept —
and names the window between the database dump and the file copy, which no
ordering closes.
Instances that accumulated project managers, favorites, board memberships
or board subscriptions pointing at a project or board that no longer exists
had no way to find or clear them.
The script reports by default and only deletes with `--apply`, which first
asks whether a backup exists and refuses outright without a terminal to ask
on. It removes nothing but reference rows: structural orphans such as a
board without a project carry user content, and a membership whose
denormalized project is stale still grants real access to an existing
board, so both are reported and left alone.
Users can pick an inactivity timeout in their preferences. A warning
appears 30 seconds before, and activity or a logout is synchronised
across open tabs.
Adds TOTP setup with QR code, login challenge, recovery codes and
trusted devices that let a browser skip the second factor for 30
days. Admins can reset another user's second factor by confirming
with their own password.
Adds optional project, board, and acting-user scope to each webhook so
events from a specific board or project can be routed to a single
endpoint without an external automation tool in the middle (closes
#1457).
A webhook now fires for an event only when every set scope matches:
empty scope still means "fire for everything", keeping existing
webhooks working unchanged. Scope filtering is centralized in the
sendWebhooks helper, which auto-derives projectId/boardId from the
event payload, so existing call sites are untouched.
Daniel Hiller