[OpenSIPS-Users] Dialog cluster syncing behaviour

[solarmon]

Hi,

I am observing an issue with the Dialog syncing between two opensips nodes,
and a corosync/pacemaker cluster setup.

I'm using 'opensipsctl fifo dlg_list' - more specifically 'opensipsctl fifo
dlg_list | grep caller_contact' - to check the Dialogs on both the 'active'
(one with the floating IP) and the 'standby' opensips nodes. Normally I can
see that these match - so I know that the clusterer module is syncing the
Dialogs normally.

However, what I'm observing is that when I take the 'standby' opensips node
down (stop the service) and bring it back up, I see that it has synced the
Dialogs that were active prior to the node going down, and it will sync any
new Dialogs since it has been back up. However, what doesn't get synced are
any new Dialogs have occurred whilst the 'standby' node has been down.

This is unexpected behaviour to me. I would expect it to sync all Dialogs
that are on the 'active' node to the 'standby' node when it has come back.

I have tried to use the command 'opensipsctl fifo dlg_list | grep
caller_contact' on both nodes, thinking that this is what is meant to be
used to force a re-sync, but that did not seem to change anything.

Can anybody confirm whether this behaviour is by design and/or whether how
to force the opensips to sync all the Dialogs between the two nodes.

Thank you for any advice you can provide.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20210625/6d10ff22/attachment-0001.html>