--Apple-Mail-8--895227565
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="us-ascii"
Hi,
I spoke too soon. While we've got it compiling, Nagios 4 still doesn't =
seem to work correctly. It appears that with the -DIOBROKER_USES_SELECT, =
the Nagios worker threads go into a loop and don't get any of the =
responses.
The test-iobroker appears to work okay, but I can't see where to look =
next. Any Solaris experts want to have a look?
Ton
On 7 Feb 2013, at 10:08, Ton Voon wrote:
> Guys,
>=20
> We've managed to fix this issue with the worker threads. It is due to =
EPOLL not being available on Solaris, so we switch to using SELECT =
instead. Patch is below (nagios_solaris_uses_select.patch), though it =
should be detected via configure instead.
>=20
> We apply various patches to Solaris builds only, some of which look =
like you've already fixed. The patches we apply are:
> * Adding include/sunos.h and base/sunos.h for missing vasprintf(), =
asprintf()
> * nagios_build_on_solaris.patch to include above - this should be =
done via configure, maybe via libnagios
> * nagios_solaris_rlimit.patch - looks like this is already applied =
by Andreas
> * nagios_solaris_compile_errors_for_comments_h.patch - this is =
probably incorrectly done but mimics Nagios 3
>=20
> Ton
>=20
>=20
>=20
>=20
>=20
>=20
> On 5 Feb 2013, at 11:45, Eric Stanley wrote:
>=20
> > Ton,
> >=20
> > I tried compiling it on my Solaris 10 (u10) x86 system and couldn't =
get=20
> > it to compile even with gcc. Do you have some patches that make it=20=
> > compile? The issues I've run into thus far include
> >=20
> > 1. the fact that RLIMIT_NPROC does not exist and
> > 2. 'typedef struct comment_struct' (core 3) in include/comments.h =
was=20
> > changed to 'typedef struct comment' in core 4 which conflicts with a=20=
> > structure of the same name in /usr/include/pwd.h.
> >=20
> > Eric
> >=20
> > On 2/4/13 4:58 PM, Gaspar, Carson wrote:
> >> I was going to try and look at this, but the current code fails to =
build with anything that isn't gcc, and I don't have the time to fix =
those bugs just to look into another one.
> >>=20
> >> Bug 1: lib/Makefile.in hard-codes gcc, and hard-codes -Wall
> >>=20
> >> Bug 2: dkhash.c fails to compile, because of an undefined variable =
i in lnag-utils.h line 111
> >>=20
> >> And that's where I gave up.
> >>=20
> >> -----Original Message-----
> >> From: Andreas Ericsson [mailto:[email protected]]
> >> Sent: Monday, February 04, 2013 8:39 AM
> >> To: Nagios Developers List
> >> Subject: Re: [Nagios-devel] Nagios 4 on Solaris 10
> >>=20
> >> On 02/04/2013 12:36 PM, Ton Voon wrote:
> >>> Hi!
> >>>=20
> >>> Has anyone tried the latest Nagios 4 code on Solaris 10?
> >>>=20
> >>> We get "(Service check did not exit properly)" for every active =
check
> >>> result. It looks to me like the worker process code has something
> >>> which is not working in Solaris. I think the communication is =
okay.
> >>>=20
> >>> There is a lib/wproc which looks like a worker test executable, =
which
> >>> executes commands from stdin via the worker process. This seems to
> >>> execute on the Solaris worker process and gets the output and =
errors,
> >>> but doesn't finish correctly.
> >>>=20
> >>> The simplest reproduction is to use the worker test executable. cd
> >>> lib && make wproc. This will fail with some missing symbols. =
Compile
> >>> by adding '-lsocket -lnsl' to the compiler flags. Run ./wproc and
> >>> enter 'hostname'. This will timeout before returning the result.
> >>>=20
> >>> Any ideas? Any Solaris users?
> >>>=20
> >> Do the self-tests pass?
> >> cd lib && make test
> >>=20
> >> ./test-iobroker and ./test-iocache are extra interesting.
> >>=20
> >> I'd be happy to help over ssh if that's an option.
> >>=20
> >=20
> >=20
> > --=20
> > Eric Stanley
> > ___
> > Developer
> > Nagios Enterprises, LLC
> > Email: [email protected]
> > Web: www.nagios.com
> >=20
> >=20
> > =
--------------------------------------------------------------------------=
----
> > Free Next-Gen Firewall Hardware Offer
> >
...[email truncated]...
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]