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.

WI03 RC and RP OLA violation report followup

From EGIWiki
Revision as of 15:21, 10 August 2012 by Mkrakowi (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



Internal procedure for COD - Availability and reliability work instruction for COD

This page describes steps which should be taken by COD shifter to follow availability/reliability issues.


When GGUS ticket about availability/reliability metrics is assigned to COD:


Timelines Step Substep Description
1 Add ticket url to Underperforming_sites_and_suspensions page
2 Ava/Rel report review
1 Prepare 'sites for suspension' list: Look at  availability metics for two previous months in AR report and the current one. If all are below 70% then sites qualifies for suspension.

Check if the site was mentioned in List of sites for which the availability followup procedures were not applicable page. In some cases there could be no need to open a ticket.

2 Prepare 'sites to be asked for explanation' list: Look at current months in AR report. If Ava. is below 70% or Rel. below 75% then sites qualifies to be asked for explanation. This list should be prepared according to requirements for input file for ticket generator.

Check if the site was mentioned in List of sites for which the availability followup procedures were not applicable page In some cases there could be no need to open a ticket.

3 Create tickets for each case as a child to the tickets assigned to COD
1 For 'sites for suspension' list please use ticket generator
2 For 'sites to be asked for explanation' list please use ticket generator
Within 10 working days from when the tickets are created. 4

Handling of sites below targets

When explanation is provided and is found satisfactory put as a solution of the ticket

'The explanation is satisfactory. Thank you!'. 

After that you should set child ticket to 'verified' status.

After 10 working days from when the tickets are created. 5 Final actions.
1 Handling of sites that are eligible for suspension
  • in the case of no NGI intervention, the site is suspended in GOC DB - as a reason put a link to GGUS ticket created for the site
  • in the case of NGI intervention, non suspension will occur if both the COD and COO agree on the reasoning provided by the NGI
    • COO should be involved to the ticket
2 Handling of sites below targets

If the explanation is not given in due time, or the explanation is found inadequate, COD send mail to NGI/ROC manager with CC to ROD and GGUS:

  • informing that NGI/ROC manager should make the site react on the ticket or suspend the site within 3 days
  • if NGI will not react COD will suspend the site on the 4th day.
Dear XX

I would like to inform you that 10 working days passed.
Please make the site react on the ticket or suspend the site within 3 days.
If NGI will not react COD will suspend the site on the 4th day.

Best Regards
XXX
On behalf of COD team
6 Prepare summary report (it should be placed in parent ticket):
  1. sites which are not responsive and didn't provided satisfactory explanation
  2. sites which were suspended
  3. ROCs/NGIs which are not responsive
  4. ...
7 Update List of sites for which the availability followup procedures were not applicable page. Put here outstanding cases which should be recorded. This could be used for example to avoid opening a ticket next month for a solved issue.
8 Update Underperforming_sites_and_suspensions page.

Questions/issues

MR: what do we do with sites marked with "n/a"?

MK: we don't take into account months with "N/A"


VERY IMPORTANT

In grid view NGIs/ROCs are named differently then in GGUS. You should change NGI/ROC name according to GGUS.


GGUS Gridview
ROC_DECH GermanySwitzerland
NGI_FRANCE NGI_France
ROC_Asia/Pacific AsiaPacific
ROC_Italy Italy
ROC_CERN CERN
ROC_Russia Russia
ROC_North NorthernEurope
ROC_UK/Ireland UKI
ROC_SE SouthEasternEurope
ROC_SW SouthWesternEurope

Tickets content

Request for explanation

Subject:$SU/$siteName - availability/reliability statistics for $date

Dear $SU,

According to recent availability/reliability report $siteName has achieved
poor performance Ava. $availability  Rel. $realiability.
More details: https://wiki.egi.eu/wiki/Availability_and_reliability_monthly_statistics.

Could you please provide explanations for poor performance of the $siteName site?

Your explanation must be returned within 10 working days from when the ticket is created.
If the explanation is not given in due time, or the explanation is found inadequate,
COD escalation procedure will be followed https://wiki.egi.eu/wiki/Operations:COD_Escalation_Procedure

If the site was certified during last month please close this ticket and 
put this info in a ticket solution field. There is known bug in report 
generation tool being worked on.


Best Regards,
EGI Central Operator on Duty

Site for suspension

Subject:$SU/$siteName site suspension

Dear $SU,

According to recent availability/reliability report $siteName has achieved
poor performance below target Ava. 50% or Rel. 50% in three consecutive months.
More details: https://wiki.egi.eu/wiki/Availability_and_reliability_monthly_statistics.

According to procedures approved on OMB 17.08, site will be suspended within 10 working days unless the NGI intervene.
If you think that the site should not be suspended please provide justification within 10 working days.

Best Regards,
EGI Central Operator on Duty

How to use ticket generator

current version of the script: 3.0

features:

  • bulk child ticket creation
  • 'assigned to' set
  • 'affected site' set
  • 'type of problem' set to Operations



  • Configure the script.

In start-explanations.pl/start-suspend.pl file at the beginning of the script you have to fill in following variable:

# PRODUCTION
my $endpoint = "https://gusiwr.fzk.de/arsys/services/ARService?server=gusiwr&webService=Grid_HelpDesk";
my $user = ""; # login to GGUS web-services
my $pass = ""; # password to GGUS web-services

# Submitter data, Those data will be used as submitter's data to create tickets
my $Mail = ""; # your email address
my $DN = "";   # your DN
my $Name = ""; # Name and Surname


  • Prepare input file.

The input plain file format for both scripts is as follow:

ROC/NGI support unit in GGUS; Site name; Availability; Reliability;

Remember that in each line should be one site and the number of semicolons should be always 4. For start-suspend.pl script Availability and Reliability values are omitted but semicolons are necessary.

example:

NGI_PL; CYFRONET_LCG2; 50%; 10%;
NGI_PL; IFJ-PAN; 15%; 3%;
  • Execute the tool

Login to machine with perl installed and execute the script as follow:

perl start-explanations.pl/start-suspend.pl PARENT_TICKET_ID "DATE" FILE_NAME

PARENT_TICKET_ID - number of "Availability/reliability statistics for *" ticket

DATE - date of the report. Format: "month year"

FILE_NAME - file with input availability/reliability data

example:

  perl start-explanations.pl 4121 "May 2010" dane.txt

Best practice

  • If the site explaining that site administrator was on holidays put as a solution "This time the explanation is found satisfactory, although for the future in case of administrators holidays site should provide administrator deputy. If it is not possible then NGI should put site which is failing in downtime. Thank you!". Close the ticket and verify it.