-
-
Notifications
You must be signed in to change notification settings - Fork 2.1k
duplicate key value violates unique constraint "rooms_pkey"
error when rejoining a room
#3374
Comments
I have the same error in my log (synapse 0.99.4) after I left Synapse fails to join the room (riot pops an "internal server error" message), and this line appears in my logs. As a result I can no longer rejoin the room. |
The same thing occurred again with
Again, this prevents me from rejoining the room (even though matrix.org homeserver seems to believe I have rejoined). Last time I "solved" it using the Should I keep this issue around to provide more precise diagnostics, and if so what kind of diagnostics would help? |
We see this error on the IRCnet bridge for the bridge user ( |
synapse v1.7.0 Results system error inviting user (and probably joining theinvite sent from the same server for a room). May be malevolent invite, though.
|
Actually my last comment has a parent error, only when first trying to join the invite:
|
If you can't join the room, that's a separate problem: please open a new issue. This issue is about the harmless |
duplicate key value violates unique constraint "rooms_pkey"
error when rejoining a room
This is intended as a precursor to storing room versions when we receive an invite over federation, but has the happy side-effect of fixing #3374 at last. In short: change the store_room with try/except to a proper upsert which updates the right columns.
This is intended as a precursor to storing room versions when we receive an invite over federation, but has the happy side-effect of fixing #3374 at last. In short: change the store_room with try/except to a proper upsert which updates the right columns.
fixed by #6968 |
This is intended as a precursor to storing room versions when we receive an invite over federation, but has the happy side-effect of fixing matrix-org#3374 at last. In short: change the store_room with try/except to a proper upsert which updates the right columns.
This happened while during an instance of #2738 (specifically #2738 (comment) )
It doesn't seem to be harmful, but errors in logs can be scary.
Version information
The text was updated successfully, but these errors were encountered: