[asterisk-dev] The New CDR system

[Vasil Kolev]

В сб, 2007-03-31 в 17:22 -0400, Steve Totaro написа:
> Vasil Kolev wrote:

.....
> > As I see it, the logic shouldn't be hard to be designed and implemented,
> > although I still haven't dug in the CDR system of asterisk, having been
> > busy with some other parts. What do you think of this?
> >   
> >   
> Isn't that what DB clustering (Single Quorum) and RAID are for? 
> 
Well, nothing guarantees you connectivity to your database, networks
also fail - and in any case, the CDR collection is not a strictly
real-time process, so there isn't any that big need for making it
real-time - CDRs should be buffered in case of connection failure and
pushed into the DB when possible. DB clustering will solve the problem
of one DB server dying (although there's almost always a period in the
fail-over when there's no database server you can access). Yes, it's
part of the solution, but not the whole solution.

> I suppose a default to write to a rotating log file like queue_log as 
> well as DB clustering would be even safe.

Yes, a rotating log should be ok, but it creates some ugly issues with
race conditions, etc. (solvable with forced rotation of the log), so
from my POV a dir with a lot of files (or in hashed subdirs for some
filesystems) is far easier to use.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: =?UTF-8?Q?=D0=A2=D0=BE=D0=B2=D0=B0?= =?UTF-8?Q?_=D0=B5?=
	=?UTF-8?Q?_=D1=86=D0=B8=D1=84=D1=80=D0=BE=D0=B2=D0=BE?=
	=?UTF-8?Q?_=D0=BF=D0=BE=D0=B4=D0=BF=D0=B8=D1=81=D0=B0=D0=BD=D0=B0?=
	=?UTF-8?Q?_=D1=87=D0=B0=D1=81=D1=82?= =?UTF-8?Q?_=D0=BE=D1=82?=
	=?UTF-8?Q?_=D0=BF=D0=B8=D1=81=D0=BC=D0=BE=D1=82=D0=BE?=
Url : http://lists.digium.com/pipermail/asterisk-dev/attachments/20070331/28e95789/attachment.pgp