> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of 
> Beeton, Carolyn (CAR:9D60)
> > > There could also be a problem in writing the file, since
> > the separate
> > > lock in each process does not guarantee that one process will not 
> > > begin to write the file while another is also in the process of 
> > > writing.
> > 
> > That is true, but the actual replacement of the file 
> contents with the 
> > new file contents is done by a rename of the temporary 
> output file to 
> > the name of the real file, and rename is implemented 
> atomically by the 
> > kernel.
> > 
> 
> I hadn't thought of that.  In that case, I think the same 
> workaround that all the other DBs use will be adequate (i.e. 
> have the supervisor preload subscription.xml and 
> registration.xml).  There is a lot of code in the various DB 
> constructors that just isn't doing what we expect 
> (initializing mTableLoaded to true, for example) - but it 
> will all be ok as long as one single process loads them 
> first.  So I think I'll proceed with the other parts of this 
> and just leave the init mess the way it is.
> 

I've taken the easy way out then and just made supervisor preload
subscription.xml.
Code review at http://code.sipfoundry.org/cru/XECS-39 if anyone wants to
comment.

Carolyn
_______________________________________________
sipx-dev mailing list [email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
sipXecs IP PBX -- http://www.sipfoundry.org/

Reply via email to