PROC09 Resource Centre Registration and Certification

Title	Site Certification Procedure
Document link	to be determined
Last modified
Policy Group Acronym
Policy Group Name	Operational Documentation
Contact Person	Vera Hansper
Document Status	DRAFT
Approved Date
Procedure Statement	A procedure for the steps involved to both register and certify new sites in the EGI infrastructure. The certification step can also be used to re-certify suspended sites.

Introduction

Certification is a pre-requisite for a Resource Centre (aka site) to become part of a Resource Infrastructure such as a National Grid Initiative (NGI) and EIRO (in Europe), or multi-country Resource Infrastructure.

This document describes the steps required

1. to register and certify a new site,
2. to re-certify a site which has been suspended.

Note: A separate document provides the process for decommissioning a site.

Through its parent Resource Infrastructure, a certified Resource Centre becomes member of the EGI Resource Infrastructure to make resources available to international user communities.

A certified site guarantees a minimum quality of service of service of these resources (currently expressed in terms of monthly availability and reliability), it must ensure troubles are handled in a timely fashion, it must understand and adhere to a common set of policies and procedures. This compares to an uncertified, or test Resource Centre, which does not provide a guarantee on the availability or usability of it's resources.

Definitions

The entities involved in this procedureare defined in the EGI Glossary.

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", “MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.

Entities involved in the procedure

• Resource Centre (or Site) Operations Manager, who is responsible of initiating the certification process by applying for membership to a Resource Infrastructure
• Resource Infrastructure Operations Manager, who is responsible of approving the integration of a new Resource Centre into the respective Infrastructure
• Operations Centre, the entity who is technically responsible of carrying out the Resource Centre certification part of the procedure, once the membership is approved

The Resource Infrastructure Operations Manager can determine with the Site Operations Manager the level of involvement of other actors.

Contact information

• EGI Operations: operations (at) mailman.egi.eu
• EGI Resource Infrastructure Providers are listed on the EGI site
• Operations Centres with the respective contact information is availanle from GOCDB

Actions and responsibilities

Site Operations Manager

1. A Resourece Infratructure Provider is responsible for all sites within the respective jurisdiction (for example, a NGI is responsible for all site in its country). For this reason, the Site Operations Manager of a new site is REQUIRED to contact the respective NGI - if the Site is located in Europe, or a Resource Infrastructure Provider active in a relevant geographical area - if the Site is outside Europe, about the intention to join the EGI infrastructure. If needed, EGI Operations can assist the Site Operations Manager to get in contact with the relevant partners (see the Contact information section).
2. The Site Operations Manager is REQUIRED to provide the necessary Site information needed to complete the registration process, is responsible of its accuracy and maintenance.
3. In order to be certified, the Site Operations Manager is responsible of reading, understanding and accepting the Resource Centre Operational Level Agreement, which defines the obligations of a Resource Centre and the committment to deliver a minimum quality of service to its future users. Endorsement of OLA implies - among other things - the acceptance of:

1. the Grid Security Policy
2. the Grid Site Operations Policy
3. the Site Registration Security Policy
4. all other policies for all EGI participants from the Security Policy Group

Resource Infrastructure Provider Operations Manager

1. A Resourece Infratructure Provider is REQUIRED to be responsible for all sites within the respective jurisdiction. For example, a NGI is responsible for all sites in its respective country.
2. The Resource Infrastructure Provider Operations Managers MUST attend site certification applications and MUST provide feedback to the requesting partners in a timely manner to accept or reject the requests received.
3. He/she MUST provide information to the Site Operations Manager about the Resource Centre OLA, and is responsible of keeping records of Site Operations Manager agreement, as deemed suitable by the Resource Infrastructure Provider (for example, through signed e-mail agreement, the collection of signatories on a paper copy of the OLA, or other means).
4. In case a request is accepted, the Resource Infrastructure Provider Operations Manager MUST contact the relevant Operations Centre to start the site registration as candidate, and the certification procedure. Registration is only needed in case of new sites.

Operations Centre

1. The Operations Centre is responsible of registering (if applicable) and of certifing the site.
2. The Operations Centre is responsible of registering an accepted site in the EGI configuration repository GOCDB.
3. The Operations Centre MUST collect the mandatory information specified by the site registration procedure, and MUST accurately input the supplied data into GOCDB.
4. The Operations Centre MUST integrate site information in all operations tools as needed, such as the local Nagios server for monitoring of uncertified sites, the local helpdesk (if available) for the registration of the Resource Centre support staff, etc.
5. In case of an existing site that is starting certification after suspension or security reasons, the Operations Centre MUST contact the EGI CSIRT to verify that all requested repair operations have been successully applied to fix the issue.
6. The Operations Centre is responsible of verifying that all tests during the 48 hour certification process are successfully passed. The Operations Centre SHALL proceed with changing the site status in GOCDB to certified only if this condition is met.

Workflow

The various steps then required by both the Resource Infrastructure Operations Manager and the Site Operations Manager are explained in the tables below. The first part for a new site is the registration process. The actual certification process, in the second table, is applicable to both new and suspended Sites.

The general status flow that a Site is allowed to follow is illustrated by the following diagram. Information on site status and on how to manipulate it available from GOCDB Documentation.

Site registration

Requirements

1. A Site MUST be part of a Resource Infrastructure and MUST be operated by an Operations Centre. If there is no suitable provider for your country, it may be that the an Operations Centre MUST first be created. A procedure exists for this, and it is documented in the Operations Centre creation procedure.

2. To satisfy Grid security requirements a Site registration procedure must capture and maintain at least the following information. The comprehensive list of required information is available (here).

• The full name of the Site.
• An abbreviated name of the Site, which must be unique within the Grid, and preferably globally unique.
• The name, email address and telephone number of the Site Operations Manager and Site Security Contact in accordance with the requirements of the Site Operations Policy.
• The email address of a managed list for contact with Site Administrators at the site.
• The email address of a managed list for contact with the site security incident response team.

3. If a Site wishes to leave the Grid or the Grid decides to remove the Site, the registration information MUST be kept by GOCDB for at least the same period defined for logging in the Traceability and Logging Policy. Personal registration information of the Site Operations Manager and Security Contact of the Site leaving the Grid MUST NOT be retained for longer than one year.

4. It is RECOMMENDED that email contacts for the Site Administrators and Security Officer(s) are mailing lists, and not individuals.

<Comment: additional contraints - if any - on information that is registered need to be specified here>

Steps

The following steps are only applicable if the Site is not already registered in GOCDB. They describe the steps for a Site Operations Manager that is requesting the respective Site to join the EGI infrastructure.

Actions falling on the NGI are the responsibility of the NGI manager. Actions falling on the Site are the responsibility of the Site Manager/Representative.

#	Responsible	Action
1	Site	Contact your NGI Manager. National contacts are available in http://www.egi.eu/production-infrastructure/Resource-providers/. Provide your NGI Manager the required information according to the template available in https://wiki.egi.eu/wiki/SiteCertMan/Required_information
2	NGI	The following actions can be done in parallel: Forward all necessary and required documentation to install and configure the site services to the Site Representative. Communicate with the Site Representative to clarify any doubts or questions. Include the NGI ROD or help-desk teams in the step if necessary.
3	NGI	Add the site to the GOCDB and flag it as "candidate". Notify the Site Representative that they should register themself in the GOCDB and request the Site Administrator role. Approve it when done.
4	Site	Complete any missing information for the site's entry in the GOCDB, including services that are to be integrated into the infrastructure. Request in the GOCDB (or ask the relevant site security staff to request) the mandatory Site Security Officer role. A security expert is the most appropriate actor for this role. Accept or deny all the requested roles under the site scope. (Caveat: If the Site manager can not approve roles, they should request the NGI manager to do so. This is a current (20.12.2010) flaw in GOCDB.) Notify the NGI Manager that the site information update is concluded.
5	Site or NGI	Check whether the site appears in the "Notified Site" field in https://gus.fzk.de/ws/ticket_search.php Note that this step should happen automatically when the site is correctly entered into the GOCDB. If this is still not visible 2 days after the GOCDB entries have been created, the NGI manager should be informed and should then contact GGUS administrators. A new site admin should register in GGUS (https://gus.fzk.de/admin/get_account.php?accounttype=support) but not specify any role, unless directed to by the NGI manager.
6	NGI	Check that the site's information is correct. (Site roles and any other additional information.) Check that contacts receive email (if they are mailing lists, check that outside EGI members are allowed to post there). Check that the required services for a site are properly registered (CE, siteBDII, SE, APEL). Check domain names and DNS.
7	NGI	Any other NGI requirements (join a certain VO, join any "ops" mailing list, etc.)
8	NGI	If all previous actions have been completed with success, notify the Site Manager that the Registration is completed.

After the successful completion of all these steps, the site is considered as to be in the "Candidate" state and is ready for the certification process.

Site certification

The Site Certification procedure is applicable for both new sites which have reached the "Candidate" state and for suspended sites, The following is a detailed description of the steps required for the transition from the "Uncertified" to the "Certified" state of the site.

#	Responsible	Action
1	Site	Notify the NGI Manager that the site is ready for certification.
2	NGI	If the site is in the "Candidate" or "Suspended" state, then flag the site as "Uncertified". If it was in the "Suspended" state then check that the reason for suspension has been cleared.
3	NGI	Check that the GIIS (gLite: BDII) is working, and publishing coherent values, namely: the correct NGI is being published in GlueSiteOtherInfo. all services registered in the GOCDB are published and ALSO that services published in the GOCDB are valid. ops and EGI dteam VOs are configured and supported. regional VOs are configured and supported. There are detailed examples for how to do this in SiteCertMan/GIIS_BDII_check.
4	NGI	Check that the registered services are fully functional by performing manual tests. e.g. from the UI or a dedicated SAM/Nagios testbed infrastructure provided by the NGI. There is an example of how to create a testbed nagios at this page. Contact the site admins if there are problems, and ensure that they fix them. Include the ROD and help-desk teams if necessary. Iterate this step with the site admins until tests pass. The prime tests to check are: network connectivity. CE job submission. SE data transfer Details for submitting manual tests can be found at SiteCertMan/Grid_manual_tests.
5	NGI	If all preliminary tests are passed for 3 consecutive days, declare an initial maintenance downtime and switch the site status to Certified. This ensures that site will appear in Nagios and GSTAT.
6	NGI	After two days check that the site appears in all operational tools. If there are problems with a specific tool, open GGUS tickets to the relevant Support Units. The major tools that are relevant are: Regional Nagios (Nagios) Operations Dashboard (Dashboard-Siteview) GridView GSTAT SAM/SFT
7	NGI	Ensure that, before the end of the maintenance downtime all Nagios tests (see above) are passed AND Accounting data is properly published. GSTAT is not in an error state. CAVEAT: There may be some problems with this tool and ARC sites.
8	NGI	Notify the Site Manager that the site is certified
9	NGI	Add site contact to any regional mailing list and provide access to any regional tool
10	NGI	(Optional?) The NGI can broadcast that a new site is now part of the EGI infrastructure.

After the successful completion of these steps, the site is considered as "Certified".

Revision history

Version	Authors	Date	Comments
0.7	Vera Hansper	2011-02-02	Updated introduction to include roles, etc. and added required documentation link for policies