Update of deployment and required packages for RHEL8?

[syuvivida]

Hello,
We are working on the upgrade of DQM machines. The physical machines are upgraded with more powerful ones and the operation system is also moving from centos 7 to RHEL8. Here we have a couple of questions and wonder if the experts could help us.

1. is there a plan to update the codes (for example the script Deploy to include the new SCRAM_ARCH el8_amd64_gcc11? Sorry I am still new to the codes here and not sure how much the SCRAM_ARCH matters here.

2. While we were trying to create a temporary DQMGUI and ran the command at our new DQMGUI servers, we find a lot of packages for bootstrap are missing. Most of them could be installed. However, one of them is not supported by RHEL8: compat-libstdc++-33

Do you know if there is a way to resolve this? Thanks!!

$PWD/deployment/Deploy -A slc7_amd64_gcc630 -r "comp=comp" -R comp@HG2209d -t MYDEV -s "prep sw post" $PWD dqmgui/bare

Shin-Shan Eiko Yu for DQM-DC

[muhammadimranfarooqi]

Hi @amaltaro @vkuznet

Could you please comment. Thanks

[amaltaro]

@syuvivida Hi Shin-Shan, I fear we do not have any central effort to move the COMP packages to a new ScramArch (be it a new OS or a new GCC version).
For the past year or two, we have actually started recommending groups to move away from the RPM-based packaging and deployment, towards the a more standard kubernetes deployment based on docker images for the services. Right now there is a mix of models in the CMS COMP system:
a) some systems managed to move completely to docker + kubernetes (no longer relying in RPMs neither in our specific COMP stack - coupled to SLC7 under gcc630)
b) other services moved to docker + kubernetes, but the docker image is still based on the RPM packages (thus, still completely dependent on the COMP stack)
c) other services have NOT been migrated to docker + kubernetes and they rely fully in the COMP stack. Many of these are legacy and/or orphan services.

I know there has been developments on a new DQM Gui and that their deployment is using kubernetes (and if I am not wrong, built together with CMSSW). So maybe the best recommendation I could give you is to discuss this with the DQM team first and see what the options are.
I let others complement and/or correct what I wrote here.

[vkuznet]

Shin-Shan, I second Alan's comments. I doubt that DQM codebase requires CC&, CC8 per-se and neither requires gcc630. You just stuck with deploy procedure used in a past. My suggestion is to reach out DQM team and evaluate the deployment procedure. One way would be to move towards docker based solution where you can build your image in a different way. This procedure can lead you to usage of k8s, or be deployed on VM. Another way is to move towards pypi installation (if your code is python based) as WMCore did. Or, you may even end-up with your custom deployment procedure. As Alan pointed out the RPM based procedure and Deploy script is outdated and we do not have any manpower to support it (which requires adjustment of scripts, making new RPM repo, building all dependencies, etc.). Therefore, a concrete plan from DQM team is required to move forward and we can't make it for you. Said that, we'll be happy to provide some additional guidance once such discussion may take place.