This is a multi-part message in MIME format.
------------7NhE3WuB9OnU6N9D2A0R5Y7
Content-Type: multipart/alternative;
boundary="----_=_NextPart_001_01C58DF8.43972774"
This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.
------_=_NextPart_001_01C58DF8.43972774
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: Quoted-Printable
Hi,
Thanks for your interest. I started to rewrite the code of this patch
because it is much too sensible to the way the compiler aligns its data
structures. I didn't try it on Linux but I tried it with several optimize
options (gcc) on Solaris and it proves that incrementing pointers in struct=
s
is not reliable.
I started rewriting the code with two dedicated serialize & unserialize
functions for each data structure, but I didn't have enough time to complet=
e
it, and I won't be available to work on it until the end of August.
Regards
Fran=C3=A7ois
> -----Original Message-----
> From: Andre Bergei [mailto:[email protected]]=20
> Sent: Wednesday, July 20, 2005 11:55 PM
> To: Laupretre, Fran=C3=A7ois (CALYON); [email protected]=
> Subject: RE: [Nagios-devel] Patch submission for comments :=20
> CGI speed improvement (XNG)
>=20
>=20
> >=09Now, the next step is to see if you find it interesting=20
> enough to=20
> >include it in a future version of
> >nagios. If you test it, please let me know how much=20
> improvement it brings in your case.
> >
> >=09Installing :
> >
> >=09The reference version for this patch is the 2.0b3 (I=20
> will do it for=20
> >the CVS version if needed). The file >names it uses are not=20
> read in configuration file yet. They must be set manually in=20
> xdata/xsdng.c and
> >xdata/xodng.c, as XSDNG_DUMP_FILE and XODNG_DUMP_FILE. In a=20
> future version, the names can be derived from the=20
> >'status_file' and 'object_cache_file' config vars, or there=20
> can be two new config vars for them, TBD.
> >
> >=09Regards
> >
> >=09Fran=E7=AF=A9s
>=20
> Hi.
>=20
> I decided to test this patch in my test enviroment against my=20
> configuration with 3000 hosts and 30000 services, but after=20
> patching nagios segfaults with
>=20
> -snip
> Starting network monitor: nagios
> /etc/init.d/nagios: line 185: 4257 Segmentation fault $NagiosBin
> -v $NagiosCfgFile >/dev/null 2>&1
> CONFIG ERROR! Start aborted. Check your Nagios configuration. -snip
>=20
> Of course, my configuration works with clean nagios 2.0 code.
>=20
> Here is my steps
>=20
> I downloaded nagios2b3
> I patched it with: cat nagios_xng_patch_1.0 | patch -p1=20
> ./configure prefix=3D/opt/nagios/ Make all (no errors) Make=20
> install (no errors)
> =20
> My system is a 2x3Ghz Xeon on Debian Sarge with GCC 3.3.5=20
> Attached is a strace dump of the nagios execution.
>=20
>=20
>=20
> cheers
> Andr=C3=A8 Bergei
>=20
------_=_NextPart_001_01C58DF8.43972774
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: Quoted-Printable
RE : [Nagios-devel] Patch submission for comments : CGI speed improv=
ement (XNG)
Hi,
Thanks for your interest. I started to rewrite the code o=
f this patch because it is much too sensible to the way the compiler aligns=
its data structures. I didn't try it on Linux but I tried it with several =
optimize options (gcc) on Solaris and it proves that incrementing pointers =
in structs is not reliable.
I started rewriting the code with two dedicated serialize=
& unserialize functions for each data structure, but I didn't have eno=
ugh time to complete it, and I won't be available to work on it until the e=
nd of August.
Regards
Fran=C3=A7ois
<FONT SI
...[email truncated]...
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]