perfsonar-dev - Re: [pS-dev] Testing Results-Important
Subject: perfsonar development work
List archive
- From: Michael Michalis <>
- To: Roman Lapacz <>
- Cc: "" <>, Roman Lapacz <>, Maciej Glowiak <>, Jason Zurawski <>
- Subject: Re: [pS-dev] Testing Results-Important
- Date: Thu, 21 Feb 2008 11:50:16 +0200
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Roman Lapacz wrote:
| Michael Michalis wrote:
| Szymon Trocha wrote:
| | Jason Zurawski pisze:
| |> Michael;
| |>
| |>> LS Registration Testing
| |>> =================
| |>> The service failed to register to LS. After investigation it was
| |>> concluded that the reason for this was incompatibility between the
| |>> generated LSRegister requests of the service and the LS implementation.
| |>
| |>
| |> I would be interested in seeing an example of the incorrect request
| |> messages that were being sent, and the responses that the LS was
| |> issuing. This should benefit all service developers and get us to a
| |> solution much faster.
| |
| Hi,
|
| I've used tethereal to capture the traffic between RRD MA and LS.(I'm attaching it). The LSRegister Request is transported in three different HTTP requests. It seems that for some reason the checksum of the of tcp packages send from the RRD MA to the LS containing the LSRegisterRequest is wrong. Just a guess, as am not sure how the HTTP reconstructs the segmented request, but it could be that the LSRegister Request parts are not reconstructed properly and thats the reason why LS registration works OK when we have small LSRegsister requests like the ones in the example (Interface_specification doc) and why the LS cannot parse bigger LSRegister requests like the ones sent by the RRD MA.
|
|
| > The problem with register request has been solved. Maciej mentioned this in his email. The problem was in LS' objects.config file. Some mappings there were missing. When Maciej updated that file sending register request (that one which didn't work in MA) started working fine.
|
:-( :-( :-[ . The funny thing is that I saw Maciejs mail that said what you have said Roman, but then I also saw Szymons mail that said to provide response and request and for some reason I started investigating. That was a nice way to start the morning....I think I need some vacations....Sorry for the spam guys.
Michalis
| > Roman
|
|
| Michalis
| | Michalis,
| |
| | Could you attach req/resp to bug #348?
| |
| | Regards,
|
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFHvUlYoEWq51/Q/40RAnpCAKCEx3ABGMxOet7m0QeKDomxue0CuQCfetR/
rjOkHLU6B31AFBmba4OiUic=
=t4/G
-----END PGP SIGNATURE-----
- Testing Results-Important, Michael Michalis, 02/20/2008
- Re: [pS-dev] Testing Results-Important, Szymon Trocha, 02/20/2008
- Re: [pS-dev] Testing Results-Important, Jason Zurawski, 02/20/2008
- Re: [pS-dev] Testing Results-Important, Maciej Glowiak, 02/20/2008
- Re: [pS-dev] Testing Results-Important, Szymon Trocha, 02/20/2008
- Re: [pS-dev] Testing Results-Important, Michael Michalis, 02/21/2008
- Re: [pS-dev] Testing Results-Important, Roman Lapacz, 02/21/2008
- Re: [pS-dev] Testing Results-Important, Michael Michalis, 02/21/2008
- Re: [pS-dev] Testing Results-Important, Roman Lapacz, 02/21/2008
- Re: [pS-dev] Testing Results-Important, Michael Michalis, 02/21/2008
Archive powered by MHonArc 2.6.16.