Hi!

    I can't give you any hints on plan A.

    On plan B I can share my experiences as I am doing more or less the 
same thing, but in a rather simpler way.

    I publish metric with probes developed in C that are then linked 
with libganglia (and some others libraries that are needed). The more 
triky parts was finding with what libraries I had to link, as well as 
how to deal with the APR memory management (as used in ganglia).

the include:
    ganglia.h

the libraries:
    ganglia apr-0 confuse aprutil-0 rt m crypt dl resolv nsl pthread

    Check to see if this libraries are instaled in your system, if not 
you can then use them from the ganglia source code package.

dealing with APR memory pools:
    this was a PITA! memory pools and glibc standard malloc/free 
management don't go together well. After trying several solutions I 
found out that if we do the initialization of the gemetric context as 
before starting to use dynamic memory then things go smooth (as far as 
I can tell).

    See the sources for gmetric.

    If you run into any problem I will glad to sent you (or anyone else) 
the sources I'm using/developing.

    Maybe someday there will be an easier and straight forward way to 
develop new metric sources for ganglia... ;)

    Hope it helps,

                  Miguel

On Thursday, 30 November 2006 17:33, [EMAIL PROTECTED] 
wrote:
> Plan B would be to snip out the gmetric code and create gmetric style
> udp
> packets for the extra metrics, but it was not clear to me that that
> would
> be OK given that the callback stuff and the xdr codes are a bit
> intertwined.

-- 

"Until they become conscious they will never rebel,
 and until after they have rebelled they cannot become conscious."

        - George Orwell's 1984 - 

Attachment: pgpE2cFHCJGAi.pgp
Description: PGP signature

Reply via email to