On Thu, 20 Jul 2006, Andrew Beekhof wrote:
> On 7/20/06, Lars Marowsky-Bree <[EMAIL PROTECTED]> wrote:
> > On 2006-07-20T14:40:20, Andrew Beekhof <[EMAIL PROTECTED]> wrote:
> >
> > > > I do believe that it is needed. Several thing require it now.
> > > but they're all supposed to be optional components right?
> >
> > Yes.
>
> cool - can the person/people that look after
> lib/plugins/stonith/apcmastersnmp.c figure out some way to allow the
> build to pass when no SNMP is available then?
As I say, it works for me routinely on Solaris (which tends to lack SNMP
and other such goodies). So I wonder whether you are seeing an effect
local to yourself (for some reason)?
The source directory has:
apcmaster.c
apcmastersnmp.c
A typical Solaris object directory here has only the "apcmaster.*o*"
variants, (i.e. not the "apcmastersnmp.*o*"). So it does successfully
detect and build non-SNMP for me.
One of my Linux object dirs is also like this; but another includes the
"apcmastersnmp.*o*" variants.
So (for me) it builds successfully in the presence or absence of SNMP.
Thus the logic would seem to be (mostly) correct. But perhaps you have
uncovered a particular instance in which it is incorrect or incomplete.
Tracing this back up:
"lib/plugins/stonith/Makefile.am" has the section:
if USE_APC_SNMP
apcmastersnmp_LIB = apcmastersnmp.la
else
apcmastersnmp_LIB =
endif
So this "USE_APC_SNMP" thing needs to be correct.
In "configure.in", this seems to be set according to "$ENABLE_SNMP".
And "$ENABLE_SNMP" is set according to the presence of various ".h" files
and a couple of other things. Admittedly its default value is an
optimistic "yes" and perhaps it might be better to invert this to a
default "no".
Look in "configure.in" around line 1000 (line: ENABLE_SNMP="yes").
Perhaps temporarily insert a "set -x" here, and a corresponding "set +x"
about 50 lines later; re-create "configure" and run it to see what route
it takes throught that section.
Chance are it will end up (for you) still with 'ENABLE_SNMP="yes"' for
some reason, when your system probably requires a "no" result. So could
you then either dream up a small additional test to do that, or perhaps
(more major) rework the logic to a default "no"?
Hope that helps, Andrew.
All the best.
--
: David Lee I.T. Service :
: Senior Systems Programmer Computer Centre :
: Durham University :
: http://www.dur.ac.uk/t.d.lee/ South Road :
: Durham DH1 3LE :
: Phone: +44 191 334 2752 U.K. :
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
