Known issues
This document provides a list of all known issues with MongooseIM operation and configuration. You may also find proposed workarounds if any are available.
Missing MUC Light room config fields with RDBMS backend¶
Before MongooseIM 3.5.x (incl.) new MUC Light rooms could be created with some config fields absent in the RDBMS table. These options couldn't be re-added later by changing the room config via requests from the clients.
It happened when the default config was a subset of the schema, and the client hasn't provided these values when a room was created.
Please note that this issue was resolved from MIM 3.6.0 onwards as the default_config
option was deleted.
How to fix this?¶
You have to iterate over all rooms in the DB (muc_light_rooms
table) and add missing entries to the muc_light_config
table.
Every option is inserted as a separate row and is stored as plain text, so it should be straightforward.
Let's say you were using the following config in mongooseim.cfg
:
1 2 3 4 5 6 7 8 9 10 |
|
Your client application has created some rooms without the background
option by mistake.
For every id
in the muc_light_rooms
table, you need to execute:
1 |
|
MSSQL connectivity via ODBC¶
We have observed some issues with he ODBC driver used by MongooseIM in the past. The problems should now be resolved, and MSSQL is verified to work on Ubuntu 20.04.2 LTS.
GDPR retrieval for MAM MUC limitation¶
When the personal data retrieval is executed for a user in a specific domain, Message Archive Management for groupchats must be running for this particular domain.
This is the case for most configurations, but the problem manifests when a MongooseIM operator configures mod_mam_muc
/mod_mam
to start only for a subset of domains supported by the cluster (host_config
option).
In such case, personal data stored by MAM MUC will not be retrieved for this user.
Proposed workaround¶
Start a dedicated MongooseIM instance with a slightly different config, which enables Message Archive Management for the user's domain. This instance doesn't have to be clustered with other nodes and doesn't have to be accessible for actual users.
After a successful retrieval, this instance may be terminated and deleted if necessary.