perfsonar development work

[Nicolas Simar <>]

Hi,

Currently, the MDM web-services are only being monitored at web-service 
layer (perfsonar ping, http-get), which enables MDM administrator to 
define the Round Trip delay, delay variation, loss, duplication, 
reachability and availability of the service. This This monitoring is 
very useful, but doesn't guarantee that

1. the user can access the functionality provided by the web-service 
(something can be broken inside a web-service and the web-service can 
still be alive) i.e. that the web-service can access the "resource 
layer", such as databases or measurement tools.

2. the service is not flapping in between service pooling interval.

Without those capabilities the MDM service cannot be offered globally 
and the SD cannot support efficiently the service. The SD needs alarm 
telling that something doesn't work anymore within the web-service.


To achieve it, we wish to create a release perfsonar 3.1 so that it can 
be used to monitor the LHC-OPN web-services.


For the web-service It is foreseen an extension of the pS echo ping. 
When sending the extended echo request, the web-service perform some 
internal checks to verify that the main resources used to deliver the 
functionality to the user can be accessed by the web-services.
The SD would use Nagios to send regular request to the web-service and 
get the results of the internal checks done by the service. If the 
checks returns a fail, the SD would receive an alarm from Nagios. 
Examples of internal checks:
                o for the BWCTL MP: the bwctl daemon is running, etc.
                o for the LS: XML Database connection: check if it is 
possible to connect to database through perfsonar service; Data store 
access: check if data store is accessible and well structured, and 
Control store access: check if control store is accessible and well 
structured


Fro Zagreb, we are asking you to identify
1) the main resources that are required by your service to offer some 
functionality to the user and
2) to identify how you can check that it is accessible/still working

You will be presenting the checks in Zagreb during the session: 
http://wiki.perfsonar.net/jra1-wiki/index.php/JRA1_Meeting_-_Zagreb#Web-service_diagnosis

You can find some examples in attachment (please read the first part of 
section 1, section 1.2 to get a general picture and look at section 3.2 
and 3.3 for some specific example if identification of the resources 
that needs to be checked).

Thank you very much in advance.

Best regards,
--
Nicolas
______________________________________________________________________

Nicolas Simar
Network Engineer

DANTE - www.dante.net

Tel - BE: +32 (0) 4 366 93 49
Tel - UK: +44 (0)1223 371 300
Mobile: +44 (0) 7740 176 883

City House, 126-130 Hills Road
Cambridge CB2 1PQ
UK
_____________________________________________________________________

Attachment: MDM-07-18v1.5 perfSONAR Service Monitoring for MDM - IL.doc
Description: MS-Word document