Skip to content
Matthew Harris edited this page Oct 9, 2013 · 13 revisions

CMIP5 Help Desk Guide

The help-desk system provides a email and

  • It is a "footprints" system, which is a commercial issue tracking package used through a web interface.
  • Queries sent to the designated helpdesk address are automatically lodged in the system.
  • Queries can be assigned assigned to appropriate people within ESGF.
  • Anyone can view the queries assigned to them and the history of interactions on that query.

Terminology

  • Agent
    • A person with a login to the footprints system who can categorise queries and respond to them.
  • Query
    • A request for help from a contact.
  • Customer
    • The user who has the query. This may be different from the submitter if the query was forwarded by an agent.
  • Contact
    • Another name for the customer.
  • Status
    • Each query has a status which controls who gets emailed when query is updated.

Anatomy of a Query

In the web interface your initial view lists your currently assigned queries. Select one of the query types in the "CMIP5 Help Desk Totals" box or from the query-type dropdown to view different query types. Incoming queries have Status=Request and are automatically changed to Status=Open when opened. Once an agent updates the query Status=Action.

When you select a query you will see the Edit Query view which shows the query information.

Standard Fields

Every query has a Title, Priority, Status and Customer listed at the top of the view.

type response
''Request'' A new query yet to be assigned - e.g. new incoming queries.
''Closed'' Used when you cannot progress the query any further. Typical situations. include: the issue has been resolved. Additional information is needed from the user before you can proceed ("the ball is in their court") or the issue cannot be resolved other issue is nothing to do with the CMIP5 help desk
''Action'' this marks queries where you are required to do some work BR (note - this could be as simple as closing the query following the user emailing in to say "thanks" !)
''Open'' this has limited use. You can use this to mark queries that highlight some additional work that could be done, but isn't essential, that you intend to do at some point. BR (do think if you are likely to actually do this, otherwise just close the query.. you could always re-open it at a later date!)
''Chasing'' If you need to follow up on something related to a query then use this status to remind yourself to do so.

Query Information

This is where internal comments and custom query attributes are stored. The information is not sent in emails and therefore not visible to the customer.

It is often useful to communicate between agents by updating the internal comments. Communication with the customer is done by updating the Description.

Assignees and Notifications

You can assign different agents to the query in this box.

The "Send Email To" section is vital for getting email notification right. Here you configure who will be emailed when you save the query. The defaults are configured in the system depending on the query Status so saving may or may not automatically email the contact and the assignees.

  • *VERY IMPORTANT: Always check the "Send Email To" field before saving a query. Will the Customer be emailed? Do you want them to be? See the screenshot at the bottom of this page.

You can also CC other people from here.

Contact Information

Footprints has an internal address book which you can fill in from this box. The name and email address is automatically synced with the contact information at the top of the query.

Description

You append new information to the query here. The full description from previous edists is not displayed until you select "View Complete Description". However, a full history of descriptions will be emailed according to the Notification options.

Initial Response after manual forwarding

If a query has been created by an agent forwarding an email from a user (e.g. if the user emailed esg-support) the query will initially have the agent as the contact. NOTE: no notification will have been sent to the user yet. The agent should enter the web app and update the query to introduce the user to the HelpDesk .

Procedure

  1. Select the query in footprints' initial view.
  2. From the template selection dropdown at the very top of the edit window select "Initial Response" and click "Use selected template". This will fill out a couple of sentences about the helpdesk in the new description box.
  3. View the complete description to get the real users name and email. Copy these into the "Contact Information" box.
  4. Insert the user's name into the new description.
  5. Read the complete description and select a "Problem type" in the query information. The query cannot be updated without setting this field. Fill out the Gateway and OpenID information if possible.
  6. Decide which agents will deal with the query and add them in the Assignees section.
  7. Insert any internal comments as necessary to communicate with the agents. Do this by clicking "Add Comment ..." and filling out the form.
  8. Add any message for the user to the new description.
  9. In "Assignees and Notifications" select "Contact" and "Assignees".
  10. Click "Save" at the bottom of the page.
  11. If in doubt about who's been emailed reopen the query and check the history. Make sure the user was emailed.

Responding by Email

If the subject line is retained, any replies to [email protected] will trigger notifications according to the default emailing rules.

In general the customer updates a query by email and Agents update via the web interface. If carfully configured, it is possible for Agents to update the query by email, however who get's notified is dependent on the default notification settings. There are various options and advantages/dissadvantages in each case:

  1. All incoming Email automatically emails the customer and agents. This has caused problems when forwarding emails into the system because esg-support was automatically set to the customer and emailed thus triggering other people's auto-responders.
  2. Agents are always emailed but customers are only emailed on certain Statuses. E.g. when Status is Action. A new query has Status=Request, therefore in this configuration you don't have the forwarding problem but the default is email the customer. When the query is initially categorised through the web interface the status is set to Action and from then on the customer is notified of any changes.
  3. Only Agents are emailed on incoming email. This avoids mistakenly emailing the customer but agents can't easily communicate with the customer by email and it's easy to forget to check the "Send Email To Customer" checkbox.

If you aren't sure whether an email has been sent visit the web interface and check the history box at the bottom of the Edit Query view. This shows who was emailed when.

Triage

Triage is the process of classifying a query and assigning the most relevant Assignees possible so that it can be dealt with as quickly as possible.

When an email arrives in the help desk one of the triage team should write an initial response to the user asking for any additional information that will help them classify the query. The query is then passed to specialist assignees. Certain common queries may be answerable by the triage team but generally triage should be a quick process that doesn't deal with detailed queries.

If this is a new request first check that the email "Submitter" is the sender of the initial email in the description. It may have been forwarded to cmip5-helpdesk by another user.

1. Determine the Gateway where the problem occurred

The ESG Federation infrastructure is composed of number of datanodes around the world (approximately 20 as of 2011-10) and a smaller number of Gateways. Each datanode publishes data to a specific Gateway then the Gateways share metadata about all the data they hold, therefore a user can discover all data from any gateway.

ESGF has agreed that each Gateway will manage queries relating to data published to their Gateway. Therefore by default a query should be assigned to the Gateway where the problem occured or the Gateway where the data of interest was published. This determination isn't always easy.

The ESGF system moves users arround between Gateways and therefore the user may not realise which Gateway they are on when the problem occurred. Sometimes the accessed Gateway can be deduced from the data describes in the query and sometimes triage will need to ask the user which Gateway they were using.

For search related queries the accessed Gateway may be different from the Gateway they finally download the data from. E.g. a user with an OpenID from BADC may browse the PCMDI Gateway to find data at NCI. If the problem is related to search it will be occuring at PCMDI but if it's related to download it is probably be occuring at NCI. If the problem is related to OpenID login it could occur at either BADC or PCMDI.

Institutes publishing CMIP5 data as of 2011/09/05

Institute Models Gateway
BCC BCC-CSM1-1 PCMDI
CCCMA CanCM4, CanESM2, CanAM4 PCMDI
CNRM-CERFACS CNRM-CM5 PCMDI
CSIRO-QCCCE CSIRO-Mk3-6-0 NCI
ICHEC EC-EARTH BADC
INM inmcm4 PCMDI
IPSL IPSL-CM5A-LR BADC
MPI-M MIP-ESM-LR DKRZ
MIROC MIROC4h, MIROC5 PCMDI
MOHC HadCM3, HadGEM2-A, HadGEM2-ES BADC
MRI MRI-CGCM3 PCMDI
NASA-GISS GISS-E2-H, GISS-E2-R PCMDI
NCC NorESM1-M PCMDI
GFDL GFDL-* PCMDI

2. Determine the User's OpenID and/or Home Gateway

Every user should have an OpenID that they received by registering with one of the ESGF Gateways. The Gateway can be deduced from the OpenID and vica- versa according to the following table.

Gateway Gateway Home Page OpenID Pattern
ESG-PCMDI http://pcmdi3.llnl.gov/esgcet/ https://pcmdi3.llnl.gov/esgcet/myopenid/*
ESG-BADC http://cmip-gw.badc.rl.ac.uk/ http://ceda.ac.uk/openid/*
ESG-DKRZ http://ipcc-ar5.dkrz.de/ http://ipcc-ar5.dkrz.de/esgcet/myopenid/*
ESG-NCAR http://www.earthsystemgrid.org/ http://www.earthsystemgrid.org/myopenid/*
ESG-NCI http://esg.nci.org.au/esgcet/ http://esg.nci.org.au/esgcet/myopenid/*
NASA-JPL http://esg-gateway.jpl.nasa.gov/ http://esg-gateway.jpl.nasa.gov/myopenid/*
ESG-ORNL http://esg2-gw.ccs.ornl.gov/ http://esg2-gw.ccs.ornl.gov/myopenid/*
ESG-NERSC http://esg.nersc.gov/esgcet/ http://esg.nersc.gov/esgcet/myopenid/*

A user may have created multiple OpenIDs to access different Gateways. If this is the case they should be encouraged to use 1 OpenID throughout. That OpenID should be the one they used to register for the CMIP5 Research group.

Experience suggests users successfuly register 1 OpenID with CMIP5 Research but may not consistenly use that OpenID. The footprints template "Multiple OpenIDs" is designed to explain this to users.

3. Deduce Problem Type

Common Queries

Query | Action User appears to lack download privileges even though they are registered for CMIP5 Research. | User is trying to use multiple OpenIDs. Use "Multiple OpenIDs" template. Missing data in Gateway | Assign to searched gateway Inconsistent search results in Gateway | Assign to all affected gateways Problem using wget scripts or MyProxy logon | Assign to Stephen Pascoe or Estani Gonzalez + affected Gateway User says his account is locked or innactive | Assign to home Gateway. they probably haven't opened the activation email Science question not specific to particular data | Assign to Martin Juckes or Karl Taylor

Principal agents for each Gateway

Gateway Agents
PCMDI Bob Drach, Michael Ganzburg
BADC Stephen Pascoe, Ag Stephens (MOHC data queries)
DKRZ Estani Gonzales
NCAR Eric Neinhouse, Nathan Hook
NCI Muhammed Atif
NASA JPL Luca Cinquini

Clone this wiki locally