- 1 Organisation
- 2 Policies
- 3 NCI Projects
- 4 Useful contacts
- 5 conda environment
Research programs (RPs) and RP leads
|Research Program||RP leads||CMS person|
|Climate Variability||Andy Hogg, Mat England, Dietmar Dommenget, Neil Holbrooke||Aidan Heerdegen, Holger Wolff|
|Heatwaves and Cold air outbreaks||Michael Reeder, Jason Evans||Paola Petrelli|
|Rainfall Extremes||Todd Lane, Julie Arblaster||Scott Wales|
|Drought||Mike Roderick, Andy Pitman||Danny Eisenberg|
Andy Pitman: Director
Todd Lane: Vice-director
We can publish blogs whenever we have something worth to be published. Just write the post and ask for a review. Note: it can be faster to publish a post from a helpdesk question right after finishing helping the user than to postpone it.
To communicate new work:
- update all documentation
- advertise on CMS news page
There is now a roster for the help desk:
|Monday||Claire / Danny|
Main tasks on "help desk day" are:
- Documentation / training update. Don't forget this page with information on editing the wiki
- Other work
We need to update our priorities regularly (~3 months). Update on Trello (and/or spreadsheet?) in consultation with the Infrastructure Committee. Communicate to CIs.
NCI Projects husbandry
- Every 6 months, send a list of members for each CoE projects to CIs. It's a reminder to clean up the memberships.
- Everyone is in charge of caring for one or several projects. See tables below.
- Need to keep track every week or so of the storage and SU usage (use ncimonitor)
- Contact users if people are on track to exceed storage
- Contact Aidan if people are on track to exceed SU.
See this page
|dt6||Andy Pitman||Land stuff|
|e14||Matthew England||Ocean stuff|
|access||Scott Wales, Claire Carouge||Software group for ACCESS, need membership to run ACCESS||Scott|
|ks32||Paola Petrelli||ARCCSS and CLEX data||Paola|
|ub4||Paola Petrelli||Era-Interim Data||Paola|
|rr3||Kate Snow||ESGF CMIP5 Australian Data Publication||Paola|
|al33||Kate Snow||ESGF CMIP5 Data Replication||Paola|
|oi10||Kate Snow||ESGF CMIP6 Data Replication||Paola|
|fs38||Kate Snow||ESGF CMIP6 Australian Data Publication||Paola|
|p66||Tony Hirst||ACCESS GCM|
|rr4||BoM ACCESS NWP data|
|wd9||Paola Petrelli||CABLE evaluation datasets||Paola|
|hh5||Claire Carouge||Climate LIEF storage grant||Paola|
|ua8||Paola Petrelli||Miscellaneous data||Paola|
|ik11||Andy Hogg||Shared COSIMA input data||Aidan|
|sx70||Claire Carouge||Shared WRF input data||Claire|
Requesting Met Office accounts for users
New accounts need a sponsor from their institution Sponsors are listed at https://code.metoffice.gov.uk/trac/admin/wiki/GroupSponsors, Current ARCCSS sponsors are Scott, Claire and Holger Instructions for an account request are at https://code.metoffice.gov.uk/trac/admin/wiki/NewUserChecklist
In brief, create a ticket at https:code.metoffice.gov.uk/trac/admin/newticket?type=account-request with the user's name and email address, like https:code.metoffice.gov.uk/trac/admin/ticket/758
|Bob Pitt||In charge of access-svn.nci.org.au. (e.g. certificate renewal etc)|
Process for update to the conda environment
Scott is responsible for the update of the stable environment every quarter and deletion of old environments. Aidan is responsible if Scott is away.
The stable environment is updated on the Tuesday after NCI's quarterly maintenance.
An announcement of the change is sent one week before the update.
CMS folks are responsible for listing highlights for the update on conda wiki page before the announcement.
If you want to ask for a delay for the update or you need to ensure the update goes as plan, or you need to retain an old environment, discuss it at a team meeting before the announcement.
CMS maintains an anaconda installation on the NCI systems. To add new packages to the unstable analysis environments clone the conda-envs repository. The master branches contains nothing but a README. The magic happens in the analysis3 branch.
Checkout the analysis branch, add the package name to environment.yml and then commit your change.
Next push the altered branch to the repository.
git push origin analysis3
Pushing the branch will trigger testing and updating on the jenkins instance. The status of the conda environment update is viewable here:
Tips and tricks if problems with conda environments
- Scott said: "for whatever reason (something to do with external c libraries I think) loading packages is sometimes order-dependent." If a package breaks the tests, put its import at the top of the list. It would then be loaded before whatever package is causing the issue.