T1.4 delivery of large data

@xgui3783 is this related to a new WP1 service? can you share with us a sufficient name for us to use as well? Will this be a DEV or PROD instance of the service you have in mind? (this will help us identify where this VM will be hosted)

DEV or PROD

PROD

sufficient name

we haven't got a name for the service yet. Tenatively, we are calling it geometry-service

It is related to

T1.4: Enabling atlas-guided dynamic data queries task desc

gitlab issue discussion the implementation of the service

hey @xgui3783 I am sorry this took me so long - thanks for the info provided.

any reasons why this service can not be deployed as a k8s based service? I am sure you know we are trying to shift to k8s based other than VMs, unless there are good reasons behind ;)

at the moment, we want to work on a proof of principle, and a quick development cycle. We feel we should be able to move quicker on a VM, even for a weak powered one, than trying to get the perfect configured CI/deployment pipeline. Sometimes, ssh into a machine, then git pull etc is simply easier.

Once it has reached a certain TRL, we will then evaluate/work on a CI pipeline that deploys it to k8s.

Of course, the alternative is to ask for some resources in k8s, use a ubuntu base image and run kubectl exec -it <podname> /bin/bash and pretend it is a VM. I just figured creating a VM on top of k8s feels like I am wasting some resources. Of course, if this is what the infra team feels is less complicated, I can instead request k8s based resource?

I think what you described with the VM / deployment / proof of principle makes absolutely sense :) If you are considering k8s on the longer run , it is good for now!

Is this VM very high priority for you at the moment? there are things that need our attention and we can work on that next week I am afraid. unless, high priority for you (thus blocking point) so we should move things around

we can wait a week or two.

Thanks for the quick response!

added Apps label

The geometry-service VM instance is created on JSCcloud.

Specs:

vCPUs: 4
RAM: 8GB
Disks:
- vda: 40GB
- vdb: 125GB

Connectivity:

External IP: 134.94.198.192
Internal IP: 172.31.4.208
Internal DNS name: geometry-service.jsccloud
HTTP
HTTPS
SSH through the jumphost( jh-1.tc.humanbrainproject.eu )

usernames:

mdussert,
npappas,
xiao,
ubuntu

VM has node-exporter running on port 7020 for metrics.

assigned to @xgui3783

@xgui3783 FYI - feel free to close this when you get all info and re open it in case you need something more on our end

added in-review label

closed

Cloud resources needed for service
OS Base (e.g. Ubuntu)	Ubuntu (>= 24.04)
# of vCPUs	2
# of vGPUs	0
RAM (GB)	8
Block storage (GB)	128GB
# of floating IPs	1

SSH keys
Name	Public key
xiao-zeph	ssh-rsa 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 xgui3783@xiao-zeph
mdussert	ssh-rsa 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 mdussert@is246197

T1.4 delivery of large data

Request for resources

EBRAINS task/component/service name

VM Requirements (Cloud Resources - Openstack)

Network

Security groups

Public-facing/HPC-facing ports and protocols

Data Requirements

Dependencies on other EBRAINS services

Other comments

Designs

Child items ...

Activity

Requirements
egress to everywhere all ports
ingress from anywhere port 80, 443, 22