Switching project of a RAM3 suite

Currently gb02 is out of compute but I want to run the regional ancillary suite on a different project

I’ve tried

  1. Switching projects in the terminal, then rose suite-run
  2. Setting project= explicitly in rose app conf
  3. Using the rose edit gui to switch the project (I’m pretty sure this is the same as 2)

But when I kick off the regional ancillary suite it still uses gb02. The last thing to try is to make a brand new persistent session on the other project. Is this what we’re meant to do? Or is there a way to get the project= setting in the rose-app conf file to be respected by the suite and I’m doing something wrong?

I’ve noted this thread suggesting changing your default project, but as I’ve already tried switchproj I don’t think this will help

Hi @ashjbarnes, I’m the triager this week and happy to help find the right expert to look into this issue.

Could you please provide a bit more detail about the environment you’re running the Regional Ancillary Suite in, as well as the relevant config files and scripts you’re using? If possible, it would also be helpful to see where the project is currently being set in the configuration. That should help us narrow down what’s causing the suite to keep using gb02.

Hi @ashjbarnes,

To switch projects with ACCESS-rAM3 you have to manually change your Gadi default project.

I think that @mlipson 's post in that thread encapsulated what you need.

I would still try that even if you had tried switchproj and it didn’t work. (i.e. I wouldn’t assume that switchproj would work due to the complexity of the suite).

Please reach out if you are still having problems.

ok I’m still having trouble. I’ve created a new persistent session called cylc-x77, started it, killed the other one and deleted all other work directories associated with the RAS that I’m trying to run

However, the suite keeps failing here, referencing my old cylc session on gb02

[FAIL] cylc run nino25th-ancil # return-code=1, stderr=
[FAIL] 
[FAIL] ERROR: No hosts currently compatible with this global configuration:
[FAIL]   suite servers -> run hosts:
[FAIL]     ['cylc-sesh.ab8992.gb02.ps.gadi.nci.org.au']
[FAIL]   suite servers -> run host select -> rank:
[FAIL]     random
[FAIL]   suite servers -> run host select -> thresholds:
[FAIL]     None

I’ve followed the ram3 instructions again but even after running this command:

cat > ~/.persistent-sessions/cylc-session <<< "<name>.${USER}.<project>.ps.gadi.nci.org.au"

my suites still have some memory of this old persistent session

Sorry I’ve I’m still missing something obvious!

edit: never mind, i just made a new terminal and reloaded rose/cylc. That seems to have gotten me running on the right session!