Ticket #115 (closed Enhancement: Duplicate)
Store failed notifications and retry them later
| Reported by: | ibc | Owned by: | support@ag-projects.com |
|---|---|---|---|
| Priority: | Minor | Milestone: | Nice to have |
| Component: | XCAP server | Version: | 1.1.0 |
| Severity: | Non-critical | Keywords: | |
| Cc: |
Description
If the PA (OpenSIPS) crashes or is properly stopped, OpenXCAP would fail when delivering notifications via MI interface. When OpenSIPS restarts it won't re-read the xcap table value if an entry already exists in "active_watchers" or "watchers". This is: the watcher won't be notified about the new state until the presentity published a new rules via XCAP.
A solution could be in OpenXCAP side:
OpenXCAP could store in a table (or memory) a list of failed MI notifications and retry them periodically (after some expiration time they are discarded).
Change History
Note: See
TracTickets for help on using
tickets.

Inaki,
This is a nice to have improvement but is much work for little added value. We use our resources for most stringent problems and features and this one in particular does not fit the profile.
Alos there is another ticket that deals with the notification between OpenXCAP and OpenSIPS which will be changed to using xcap-diff Event package.