-
Notifications
You must be signed in to change notification settings - Fork 31
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add IBM Cloud IKS docs #49
base: main
Are you sure you want to change the base?
Conversation
$ ibmcloud ks cluster create classic \ | ||
--name <CLUSTER_NAME> \ | ||
--zone dal10 \ | ||
--flavor gx2-8x64x1v100 \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I am unable to find any gx2
flavor instances, are these still available? Do we need to enable something specific on our account.
cc @marifse
$ ibmcloud ks flavors --zone dal10
OK
For more information about these flavors, see 'https://ibm.biz/flavors'
Name Cores Memory Network Speed OS Server Type Storage Secondary Storage Flavor Class Provider
b3c.16x64 16 64GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
b3c.16x64.300gb 16 64GB 1000Mbps UBUNTU_18_64 virtual 25GB 300GB - classic
b3c.32x128 32 128GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
b3c.4x16 4 16GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
b3c.56x242 56 242GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
b3c.8x32 8 32GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
c3c.16x16 16 16GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
c3c.16x32 16 32GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
c3c.32x32 32 32GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
c3c.32x64 32 64GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
m3c.16x128 16 128GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
m3c.30x240 30 240GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
m3c.48x384 48 384GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
m3c.4x32 4 32GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
m3c.56x448 56 448GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
m3c.64x512 64 512GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
m3c.8x64 8 64GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
mb4c.20x192 20 192GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
mb4c.20x384 20 384GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
mb4c.20x64 20 64GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
mb4c.20x64.2x1.9tb.ssd 20 64GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
mb4c.32x384.3.8tb.ssd 32 384GB 10000Mbps UBUNTU_18_64 physical 2TB 1.92TB - classic
mb4c.32x384.6x3.8tb.ssd 32 384GB 10000Mbps UBUNTU_18_64 physical 2TB 1.92TB - classic
mb4c.32x768.3.8tb.ssd 32 768GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
mb4c.48x1536 48 1536GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
me4c.4x32 4 32GB 10000Mbps UBUNTU_18_64 physical 2TB 2TB - classic
me4c.4x32.1.9tb.ssd 4 32GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
mg4c.32x384.2xp100 32 384GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
mg4c.48x384.2xv100 48 384GB 10000Mbps UBUNTU_18_64 physical 2TB 960GB - classic
u3c.2x4† 2 4GB 1000Mbps UBUNTU_18_64 virtual 25GB 100GB - classic
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
hey @jacobtomlinson! thank you four response. To provision IBM Kubernetes cluster, you can use the GPU enabled mg4c.32x384.2xp100 (bare-metal server) instance, though it will incur charges per month one time not hourly, so would be careful. And for your information the gx2-8x64x1v100 is only available for Virtual Server Instance, which is a virtual instance and charged per hour basis. Thank you
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the response @marifse.
I'm apprehensive about providing instructions on deploying a very expensive cluster that bills monthly 😅. Your original instructions used gx2-8x64x1v100
which I've just copied over here, is there any way to get per-hour billing for GPUs on IKS?
We also need to test these instructions for every RAPIDS release (every two months) and right now this means launching two nodes and checking that everything runs. We can't really justify ~$1k per test because of a limitation around the billing period.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@marifse gentle nudge here, do you have any suggestions on how we can safely document IKS?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jacobtomlinson sorry for being late, i have communicated to IBM guy, for now there is no option in IKS (IBM Kubernetes Services) to use GPU enabled instance without bare-metal profile.
For testing the instructions for every RAPIDS release, i have communicated to IBM guy, he will sort out some solution for this, as soon as he will be back to work, as he is on leave for next 15 days.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @marifse I'll leave this as a draft in the meantime, but the single-node docs are already merged and will be in the next stable release which we hope to get out this week.
Follow up to #37.
Adds IKS documentation submitted by @marifse.