[asterisk-users] Write Asterisk CDR MySQL records to multiple servers

[Tilghman Lesher]

On Thursday 11 September 2008 15:47:19 Eric Chamberlain wrote:
> On Sep 10, 2008, at 7:11 PM, Tilghman Lesher wrote:
> > On Wednesday 10 September 2008 19:55:15 Eric Chamberlain wrote:
> >> On Sep 10, 2008, at 2:01 PM, Tilghman Lesher wrote:
> >>> On Wednesday 10 September 2008 13:22:51 Ricardo Melendez wrote:
> >>>> Hi to all, I actually have an asterisk server configured to write
> >>>> CDR mysql records in the same machine (localhost), but I want
> >>>> to write this records to another machine also in mysql at the
> >>>> same time, It is possible? 
> >>>> It means that I want save the records in both machines.
> >>>
> >>> You can either use MySQL replication or you can use 2 different
> >>> CDR drivers at the same time, such as ODBC, with
> >>> the Mysql-ODBC-Connector and the MySQL CDR
> >>> driver.  Also, in 1.6, cdr_adaptive_odbc allows you to
> >>> specify multiple CDR backends within the same
> >>> configuration file. 
> >>
> >> Are there any sample config's explaining how to setup
> >> cdr_adaptive_odbc?  Is cdr_adaptive_odbc used with cdr.conf?
> >> cdr.conf makes no mention of how to use cdr_adaptive_odbc,
> >> yet cdr_adaptive_odbc.com doesn't explain how to set things
> >> like usegmtime or loguniqueid.
> >
> > I thought that the sample cdr_adaptive_odbc.conf was rather clear,
> > but apparently not.  The point of this module is to allow you
> > to log whatever you like in terms of the CDR variables.  Do you
> > want to log uniqueid?  Then simply ensure that your table has
> > that column.  If you don't want the column, ensure that it does
> > not exist in the table structure.  If you'd like to call uniqueid
> > something else in your table, simply provide an alias in the
> > configuration file that maps the standard CDR field name
> > (uniqueid) to whatever column name you like.  Perhaps you'd
> > like some extra CDR values logged that aren't in the standard
> > repertoire of CDR variables (some that come to mind are
> > certain values used for LCR:  route, per_minute_cost, and
> > per_minute_price).
> > Simply set those CDR variables in your dialplan, i.e.
> > Set(CDR(route)=27), ensure that a corresponding column
> > exists in your table, and cdr_adaptive_odbc will do the rest.
> >
> > I do agree that I have overlooked gmtime as a possible setting
> > for datetime fields in cdr_adaptive_odbc, and that's probably
> > something that I need to add.  However, I think that the method
> > by which specifying which columns you'd like to have is certainly
> > much more intuitive than the old "These are your columns.
> > You must have them, or all CDRs will fail" approach.  Hopefully,
> > you will also see what I think is a rather innovative approach to
> > CDRs and wonder how you ever got along without it.  In fact, the
> > adaptive approach has been now ported to most of the other
> > CDR drivers, including mysql, postgres, and sqlite, and another
> > developer (who is more familiar with that API) is working on 
> > TDS support.
>
> Thanks for the explanation, it is clear now.  The confusing part for a
> first time ODBC user is weeding all the old unneeded cruft.
>
> Does cdr_adaptive_odbc have any support for spooling records while a
> database connection may be down?

It does not, no, but another developer is working on such changes at
a different level, which should allow all CDR backend drivers to take
advantage of the same spooling facility.

-- 
Tilghman