Alert.png The wiki is deprecated and due to be decommissioned by the end of September 2022.
The content is being migrated to other supports, new updates will be ignored and lost.
If needed you can get in touch with EGI SDIS team using operations @ egi.eu.

PROC10 Recomputation of SAM results or availability reliability statistics

From EGIWiki
Revision as of 11:38, 18 August 2014 by Krakow (talk | contribs)
Jump to navigation Jump to search
Main EGI.eu operations services Support Documentation Tools Activities Performance Technology Catch-all Services Resource Allocation Security


Documentation menu: Home Manuals Procedures Training Other Contact For: VO managers Administrators


Title Recomputation of SAM results and/or availability/reliability statistics
Document link https://wiki.egi.eu/wiki/PROC10
Last modified 1.2
Policy Group Acronym OMB
Policy Group Name Operations Management Board
Contact Group operations-support@mailman.egi.eu
Document Status Approved
Approved Date 26 March 2012
Procedure Statement This procedure documents the steps for requesting a correction in the SAM test results and in the related availability/reliability statistics.
Owner Owner of procedure



Overview

This procedure documents the steps for requesting a correction in the OPS VO SAM test results and in the related availability/reliability statistics if applicable.

DISCLAIMER: This procedure is only applicable to EGI OPS test results. Procedures for the computation of VO-specific availability report are VO-specific and are out of this scope.

Definitions

Please refer to the EGI Glossary for the definitions of the terms used in this procedure.

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.

Who can submit a request?

Re-computations can be requested by:

  • site administrators
  • regional operations staff.

Re-computation policy

Starting from the 01 May 2012 monitoring results can be recomputed only in the case of problems with the monitoring infrastructure itself. No re-computations will be performed in case of issues with the deployed middleware (e.g. in case of documented bugs affecting the availability of a production service end-point), which will be consequently reflected in lower availability/reliability.

Some examples of possible issues justifying a re-computation request:

  • invalid proxy certificate used for submitting the monitoring probes in a Nagios instance;
  • problems with the Storage Element used for replica management tests resulting in errors on CE's metrics.

The deadline for requesting re-computations is 10 calendar days after the publication and announcement of the monthly Availability/Reliability reports for a given month X (typically the announcement will be distributed on the 1st day of month X+1).

According to the re-computation requests received, A/R reports will be regenerated only once for each month, after the 10th of month X+1.

How to request a re-computation of OPS monitoring results

The request is originated by a site

STEP 1 (RC). As soon as the problem is detected, notify your NGI operations centre by opening a GGUS ticket. Please address the ticket to your Operations Centre support unit, who is responsible of validating the request. In the GGUS ticket you must mention:

  1. the starting and ending time of the problem (including day and hour in UTC)
  2. the Site, NGI/federation of NGIs affected by the problem
  3. the VO affected by the problem (must be the OPS VO)
  4. a description of the problem

STEP 2 (OC). The NGI operations centre validates the request.

STEP 3 (OC). If the request is deemed valid, a GGUS ticket is sent to Service Level Management(SLM) Support Unit. The SLM support team will take care of discussing all requests received with the SAM team.

STEP 4 (SLM SU). The SLM SU is responsible of

  1. validating the reported problems
  2. discuss the reported problems with the SAM Support Unit if needed
  3. notify the SAM SU about the requests received through a new parent ticket is submitted to SAM with the children tickets of the validated requests

STEP 5 (SAM SU).

  1. The SAM Support Unit is responsible of checking the requests and of regenerating the results. For the accepted requests all Nagios metric results for any site and service are set to unknown status from the beginning of the hour reported in the starting time to one hour after the ending time. This is to cover late results that could have arrived later. the availability and reliability of other sites won't be affected, as unknown periods are not considered in the computation.
  2. New monthly availability statistics will be recomputed for that particular period, Site, NGI/federation of NGIs.
  3. A new report will be made available 10 days after the first publication of the report.
  4. After publication of the new report, all child GGUS tickets will be closed.

STEP 6 (SLM SU).

  1. The parent ticket is closed.

The request is originated by an NGI/EIRO operations centre

Follow the procedure defined in the section above, starting from STEP 3.


External links

Tips

  • Date formats

You can use the Unix date command to convert the start and end time from your time zone to UTC using the ISO 8601 format.

the start time must be rounded to the lower hour and the end time rounded to the higher hour

Example:

# date --date="12 Feb 2012 17:00 CET" --utc --iso-8601=hours

will give:

2012-02-12T16:00+0000

Revision history

Version Authors Date Comments
George Fergadis/AUTH 03/05/2012 updated policy and procedure to reflect the OMB decision of the March 2012 meeting
George Fergadis/AUTH 17/01/2012 the text of the procedure is fixed to clarify that both RC administrators and regional operations staff can request a re-computation.
George Fergadis/AUTH 16/01/2012 the text of the procedure is fixed to clarify that the recomputation of test results can be requested before the end of the affected month, in which case if sufficient time is allowed for fixing of the test results, no re-computation of availability/reliability statistics will be needed.