Hello,
I think I’ve found the issue, so please don’t investigate this for now.
I hadn’t properly followed the step below.
Thanks,
Andrew
Optional configuration settings
To aid portability, the suites use site-specific optional configuration settings which override default values for some parameters when suite conf → Machine Options → Site at which model is being run is set to Archer2. In particular, when running on ARCHER2, the names for ancillaries in
um → namelist → Reconfiguration and Ancillary Control → Configure ancils and initialise dump fields
are specified in app/um/opt/rose-app-archer2.conf.
I’m afraid that I’m still getting the same error. I added the ancils below (specified in app/um/opt/rose-app-archer2.conf) and also set the ‘source’ as ‘initialise from ancillary file’.
You don’t need to add the ancils manually; that section was just alerting you to the fact that they are in an optional configuration file that gets automatically loaded when you set the “Site at which the model is being run” to Archer2.
I ran the reconfiguration but it fails with the error below, which seems to be a permission error related to reading in the start file. The file is obviously there, and the permissions seem correct. So I’m a little confused what the issue is.
Thanks,
Andrew
OPEN: File Open Failed: Permission denied
???
??? WARNING ???
? Warning code: -12
? Warning from routine: io:file_open
? Warning message: An error occurred opening a file
? Warning from processor: 0
? Warning number: 2
???
???
???!!!???!!!???!!!???!!!???!!! ERROR ???!!!???!!!???!!!???!!!???!!!
? Error code: 1
? Error from routine: io:file_open
? Error message: Failed to open file /work/n02/n02/anmcr/cylc-run/u-cs003/share/data/cs003a.astart
? Error from processor: 0
? Error number: 3
???
The suite u-cj514 runs straight out of the box with just the changes detailed in the above link and I suggest turning off the mirad of development tests which you don’t need. I ran it ok last night.
Can I suggest you try taking a clean copy of the that suite and run again please?
I started from scratch, and also switched off all the ‘tests’. However, I get the same error as above, which seems to be around not having the correct permission to open my reconfigured start file. See below.
Best wishes,
Andrew
OPEN: File Open Failed: Permission denied
???
??? WARNING ???
? Warning code: -12
? Warning from routine: io:file_open
? Warning message: An error occurred opening a file
? Warning from processor: 0
? Warning number: 2
???
???
???!!!???!!!???!!!???!!!???!!! ERROR ???!!!???!!!???!!!???!!!???!!!
? Error code: 1
? Error from routine: io:file_open
? Error message: Failed to open file /work/n02/n02/anmcr/cylc-run/u-cs043/share/data/cs043a.astart
? Error from processor: 0
? Error number: 3
???
I hadn’t appreciated you had run the suite several times, with it working fine on the first attempt and then failing subsequently. I can now see what’s going on.
1st run with recon - ainitial symlink to JWs start file, reconfigured dump written to astart
2nd run without recon - astart symlink to JWs start file - fails as dump needs reconfiguring
3rd run with recon - ainitial symlink to JWs start file, can’t write reconfigured dump as astart previously symlink’d to JW in 2nd run; permission denied.
This is a problem with the suite setup.
In app/install_cold/opt remove the file rose-app-no-recon.conf.
This will mean when reconfiguration is switched off after having run it successfully once, the reconfigured dump will be used rather than defaulting back to the ainitial file.
You’ll need to do a rose suite-run --new to clear things out.
I will let the Met Office know so that they can fix the suite.
I noticed that you’ve taken a copy of the suite branch u-cj514/porting_fixes rather than the trunk. I think it is probably ok but there are a couple of edits in the trunk that aren’t in the porting_fixes branch. So to be safe I would suggest switching to using the trunk. Always take the suite trunk unless instructions explicitly say otherwise.
Thanks for your quick reply. I implemented all the suggestions you made, and the model is running successfully (it hasn’t finished yet, but I can see that its moving forward through the model timesteps).
