42 dev config #43
42 dev config #43
Conversation
|
The model version in the
|
|
@blimlim arrggghhh - why is this failing? its just a straight up copy of SpinUp_b which has always worked, including today. |
|
Hi @JhanSrbinovsky! The deployment system has different requirements depending on whether a pull request is marked as a draft or not, which I think is bringing up this error. When a PR is not a draft, the deployment assumes that you're making a new release or new version of the model. In this case it requires the version number in specs:
- [email protected]_2024.12.0
...
projections:
access-esm1p6: '{name}/dev_2024.12.0'to be updated to something new. For pull requests marked as drafts, this requirement doesn't apply and it will happily deploy prerelease executables without changes to the version number. Is the goal to merge this PR into Otherwise if you're building some executables for testing, marking this PR as a draft should then let the build go ahead. |
|
The model version in the
|
|
!redeploy |
|
🚀 Deploying access-esm1.6 Details and usage instructionsThis
This Prerelease is accessible on Gadi using: module use /g/data/vk83/prerelease/modules
module load access-esm1p6/pr43-1where the binaries shall be on your 🛠️ Using: spack |
|
@blimlim - looks like the same error |
|
Hi @JhanSrbinovsky, I think the deployment itself was successful despite the messages from github! I've been able to load the modules listed: I think the |
|
Yeah. The failing check is from the initial deployment of the commit. The The reason you're still seeing the red cross is because GitHub treats previous failed checks as a higher priority than new successes...bit of a quirk of GitHub! |
|
Thanks @CodeGat! |
|
Redeploying is quite simple thanks to @CodeGat! You just have to add a comment saying "!redeploy" |
|
Indeed. Its main use case was for redeploying changes to a |
|
🚀 Deploying access-esm1.6 Details and usage instructionsThis
This Prerelease is accessible on Gadi using: module use /g/data/vk83/prerelease/modules
module load access-esm1p6/pr43-2where the binaries shall be on your 🛠️ Using: spack |
JhanSrbinovsky
temporarily deployed
to
Gadi Prerelease
GitHub Actions
Inactive
|
!redeploy |
|
🚀 Deploying access-esm1.6 Details and usage instructionsThis
This Prerelease is accessible on Gadi using: module use /g/data/vk83/prerelease/modules
module load access-esm1p6/pr43-3where the binaries shall be on your 🛠️ Using: spack |
|
🚀 Deploying access-esm1.6 Details and usage instructionsThis
This Prerelease is accessible on Gadi using: module use /g/data/vk83/prerelease/modules
module load access-esm1p6/pr43-4where the binaries shall be on your 🛠️ Using: spack |
JhanSrbinovsky
temporarily deployed
to
Gadi Prerelease
GitHub Actions
Inactive
|
🚀 Deploying access-esm1.6 Details and usage instructionsThis
This Prerelease is accessible on Gadi using: module use /g/data/vk83/prerelease/modules
module load access-esm1p6/pr43-5where the binaries shall be on your 🛠️ Using: spack |
JhanSrbinovsky
temporarily deployed
to
Gadi Prerelease
GitHub Actions
Inactive
JhanSrbinovsky
temporarily deployed
to
Gadi Prerelease
GitHub Actions
Inactive
|
🚀 Deploying access-esm1.6 Details and usage instructionsThis
This Prerelease is accessible on Gadi using: module use /g/data/vk83/prerelease/modules
module load access-esm1p6/pr43-6where the binaries shall be on your 🛠️ Using: spack |
|
Active runs from dev-access-esm1,6 with varying runtime configs:
|
|
"LUH3 restart" output is AT /scratch/p66/jxs599/access-esm/archive/dev-LUH3-restart-expt-5eed0c42 |
|
🚀 Deploying access-esm1.6 Details and usage instructionsThis
This Prerelease is accessible on Gadi using: module use /g/data/vk83/prerelease/modules
module load access-esm1p6/pr43-7where the binaries shall be on your 🛠️ Using: spack |
|
@JhanSrbinovsky - yeah, it is a bit annoying having to rebase due to a conflict when there is no intention to merge...unfortunately that's a restriction set by GitHub, where they don't run checks (and therefore deployments) if there is a merge conflict. I can have a look at the spack logs to see what is being built - which commits/how many commits back do you want me to check? |
|
@blimlim Can you please let me know when this happens. The "conflict" on our end cant be resolved so we've decided to run side by side simulations (at least for a few days). Thanks |
|
Hi @JhanSrbinovsky, the deployment after resolving the conflict is available here: #43 (comment) |
|
I've just learnt that @blimlim (who was helping me co-ordinate these things) is away today @CodeGat - maybe we'll worry about checking the logs later. I honestly dont think there's a problem there. We are going to run comparable side-by-side simulations which should answer that question anyway. On the build side we wanted to be incorporating Pears'slatest ocean, Dave Bi's CICE update. I'm not sure of the status of the spack build and what I need to do on my end? We were hoping to use the new sapphire (@manodeep was involved with this) rapids config. which I thought was being added to the default config? But when I payu clone-d yesterday I didnt pick up any notmalsr changes or anything? |
Thanks @blimlim - I just heard that you were away? |
JhanSrbinovsky
temporarily deployed
to
Gadi Prerelease
GitHub Actions
Inactive
|
🚀 Attempted to deploy 🖥️
|
|
oh no - how did a conflict re-emerge. Following @blimlim fixing the conflict, I git pull-ed the latest into my local copy. no conflicts. changed the commit hash/ ci/push and now there is a conflict again. |
|
It's because there have been new changes to ADDED: In case it wasn't clear: you didn't do anything wrong. |
|
yes it is for testing. Im fine with resolving the conflicts but the problem is that the differences are outside of the UM7 confines of my domain and not lines that I have ever modified. I didnt rebase from main either so IDK why the other side of the conflict is from main. At any rate main seems to have a later tag for cice or whatever which would be what I want - I think. I basically want to pick up the same Ocean, WOMBAT , tracers, ice etc that Pearse is using. I will update the UM7 location and continue a run from where he is now. |
|
The warning is just telling you that there are conflicts between this branch and @pearseb could you please point us to the prerelease PR that you are using so that we can makes sure @JhanSrbinovsky has the right versions for non-UM components? |
|
OK thanks Dougie - makes sense. |
|
OK so 👍 modules: So everything I need is here: |
|
it looks like my spack.yaml is already ok. I am now using (updated by @blimlim it must be) a commit hash to cice4 but I suspect this is actually DaveBi's branch head?" I cant see cice4 to verify |
Yup that's correct. |
JhanSrbinovsky
temporarily deployed
to
Gadi Prerelease
GitHub Actions
Inactive
|
🚀 Attempted to deploy 🖥️
|
|
Thanks Dougie - do you know about the saphirerapids payu config? |
JhanSrbinovsky
temporarily deployed
to
Gadi Prerelease
GitHub Actions
Inactive
|
its OK - I think ive picked up the branch directly. It isnt merged to anywhere yet, but I think/hope it will work from here. comments suggest that there's a good chance :) |
|
🚀 Attempted to deploy 🖥️
|
|
@JhanSrbinovsky Could you figure out the changes necessary to run on the sapphirerapids queue? |
|
@manodeep I think so. appears to be humming along anyway payu clone https://github.com/ACCESS-NRI/access-esm1.6-configs -B preindustrial-sapphire-rapids Actually there were some extra things I added/kept from my old config.yaml |
|
🚀 Attempted to deploy 🖥️
|
see issue #42
🚀 The latest prerelease
access-esm1p6/pr43-18at 94083d9 is here: #43 (comment) 🚀