That's excellent news, Andreas! I'm not able to see immediately how the two can work nicely together, but I'm sure it should be possible to get SER's FIFO net-centric now ;-)
For some reason, however, I think an alternative to SIP-based replication is somewhere down the line. Paul relies on mysql two-way replication and I follow his reasoning that replication should be done at the DB layer (which we do as well, but using LDAP. The usrloc replication is still SIP-based...). However, the solution where forward_tcp() and save_memory() are used, you get both SIP replication traffic, as well as DB replication. If one has full TCP/UDP-based FIFO access, shouldn't it be easy to write a small module that will send a ul_add command to another server? I wonder why Paul and you haven't done that yet, but still rely on SIP-based updates?
g-)
Andreas Granig wrote:
Hi,
Greger V. Teigre wrote:
Then, a new replication module should use XMLRPC or whatever protocol the patch is using to forward (in a guaranteed mode) the usrloc information
Apropos XMLRPC: my boss accepted the release of our XMLRPC-server under the GPL, so I'll post a link as soon as I get a chance to write some install notes for it.
It currently works as standalone daemon communicating with SER by FIFO commands via unix sockets. But now since Paul released a TCP/UDP-patch for the FIFO-system, maybe one could extract the xmlrpc parts into a SER module.
Andy