Re: [Nagios-devel] gSOAP communication between modules/plugins
Posted: Fri Apr 27, 2007 10:38 am
--nextPart2979415.lAAktPANNy
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Hi Sebastian,
On Friday 27 April 2007 14:23, Sebastian Ganame wrote:
[...]
> However, WS-MAN (along with WS-Notification, WS-Eventing, etc.) are taking
> the lead when it comes to Datacenter manageability. In the near future,
> most of the devices currently implementing SNMP traps for alerting (just =
as
> an example) will also support WS-MAN for device management, and maybe
> having the alternative to get there is one area that Nagios might/should
> investigate.
Yes, WS-Man looks pretty interesting.
Are you thinking this is something Nagios should be able to monitor (as in=
=20
some kind of check_wsman plugin), or that Nagios' current status should=20
somehow be made available via WS-Man?
BTW, out of curiosity, do you know of any kit that supports WS-Man just now=
? =20
There's a few software solutions out there, so I guess its possible for=20
existing software to be instrumented with WS-Man. Looks like Win. Server=20
2003 supports it (and some Sun hardware) but I was wondering if anything el=
se=20
does.
> I understand however that adding XML based protocols like Web=20
> Services for data gathering might impact Nagios performance, since both X=
ML
> parsing (very time-consuming) as well as increased network traffic might
> affect Nagios performance. I really would like to listen to ideas and or
> opinions regarding this.
Well, from what I've heard, the killer for short interaction WS-based servi=
ces=20
is actually the crypto stuff (e.g. validating the TLS chain of trust) rathe=
r=20
than serialising/deserialising the XML. I guess if one can keep the=20
connection live (via HTTP/1.1 persistent connections), this can be mitigate=
d,=20
but then you need something (Nagios-side) that stays alive between successi=
ve=20
probes.
If the probes are genuinely stateless, then RFC 4507 has some ideas about h=
ow=20
best to go about caching the connection-credentials, but I guess it depends=
=20
largely on what the server supports.
Of course, this isn't a problem if there's no security. But since WS-Man=20
gives you control of the device, is anyone implementing it without at least=
=20
some level of TLS support (e.g. with just Diffie-Hellman)?
Cheers,
Paul.
--nextPart2979415.lAAktPANNy
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQBGMkMICvCDPV5t1VQRAix4AKCn0Bn1SkrnhzJzi9yCYoXdsRigPQCgs062
J78zbBl2kNmEOOzm0yE5XaE=
=VT4h
-----END PGP SIGNATURE-----
--nextPart2979415.lAAktPANNy--
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Hi Sebastian,
On Friday 27 April 2007 14:23, Sebastian Ganame wrote:
[...]
> However, WS-MAN (along with WS-Notification, WS-Eventing, etc.) are taking
> the lead when it comes to Datacenter manageability. In the near future,
> most of the devices currently implementing SNMP traps for alerting (just =
as
> an example) will also support WS-MAN for device management, and maybe
> having the alternative to get there is one area that Nagios might/should
> investigate.
Yes, WS-Man looks pretty interesting.
Are you thinking this is something Nagios should be able to monitor (as in=
=20
some kind of check_wsman plugin), or that Nagios' current status should=20
somehow be made available via WS-Man?
BTW, out of curiosity, do you know of any kit that supports WS-Man just now=
? =20
There's a few software solutions out there, so I guess its possible for=20
existing software to be instrumented with WS-Man. Looks like Win. Server=20
2003 supports it (and some Sun hardware) but I was wondering if anything el=
se=20
does.
> I understand however that adding XML based protocols like Web=20
> Services for data gathering might impact Nagios performance, since both X=
ML
> parsing (very time-consuming) as well as increased network traffic might
> affect Nagios performance. I really would like to listen to ideas and or
> opinions regarding this.
Well, from what I've heard, the killer for short interaction WS-based servi=
ces=20
is actually the crypto stuff (e.g. validating the TLS chain of trust) rathe=
r=20
than serialising/deserialising the XML. I guess if one can keep the=20
connection live (via HTTP/1.1 persistent connections), this can be mitigate=
d,=20
but then you need something (Nagios-side) that stays alive between successi=
ve=20
probes.
If the probes are genuinely stateless, then RFC 4507 has some ideas about h=
ow=20
best to go about caching the connection-credentials, but I guess it depends=
=20
largely on what the server supports.
Of course, this isn't a problem if there's no security. But since WS-Man=20
gives you control of the device, is anyone implementing it without at least=
=20
some level of TLS support (e.g. with just Diffie-Hellman)?
Cheers,
Paul.
--nextPart2979415.lAAktPANNy
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQBGMkMICvCDPV5t1VQRAix4AKCn0Bn1SkrnhzJzi9yCYoXdsRigPQCgs062
J78zbBl2kNmEOOzm0yE5XaE=
=VT4h
-----END PGP SIGNATURE-----
--nextPart2979415.lAAktPANNy--
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]