Skip to content

issues Search Results · repo:openshift/instaslice-operator language:Go

Filter by

112 results
 (69 ms)

112 results

inopenshift/instaslice-operator (press backspace or delete to remove)

Instaslice Max capacity for a profile should only be added by considering the correct number of MIG-enabled GPUs on the node.
tech-preview
  • asm582
  • Opened 
    2 days ago
  • #475

For clusters with heterogeneous GPUs, we see daemonset spawned on nodes that do not have MIG-enabled GPUs; a fix should be on the node selectors in the daemonset code.
  • asm582
  • Opened 
    4 days ago
  • #474

When a node has mixed GPUs (mig enabled and mig disabled) we see incorrect node GPUs: Node Gpus: Gpu Memory: 0 Gpu Name: Gpu Uuid: Gpu Memory: 0 Gpu Name: Gpu Uuid: ...
  • asm582
  • Opened 
    4 days ago
  • #473

InstaSlice creates an instance of CR for every node that has GPU(s). it should only create instances of CR for nodes that have GPU and MIG mode enabled.
tech-preview
  • asm582
  • Opened 
    9 days ago
  • #466

GPU operator has a feature to enable and disable MIG mode. MIG gets disabled on the select node(s), but extended resources continue to exist, there is a need to design a strategy to remove extended resources ...
  • asm582
  • Opened 
    9 days ago
  • #465

We test that we do not ungate more pods than we can fit on available gpus by launching 8 long-running 1g pods and checking exactly 7 are running (in a single gpu setup). We should extend such a test to: ...
tech-preview
  • tardieu
  • 2
  • Opened 
    13 days ago
  • #458

In my testing I observed that the instaslice objects were created (and they were missing status) for the nodes that have the relevant GPUs but those GPUs do not have the MIG mode enabled. For such cases, ...
  • harche
  • Opened 
    17 days ago
  • #452

Add hierarchy to Node Gpus as such and move under Spec, Node Resources: Current: Status: Node Resources: Node Gpus: Gpu Memory: 40Gi Gpu Name: NVIDIA A100-PCIE-40GB Gpu Uuid: ...
  • mamy-CS
  • Opened 
    17 days ago
  • #450

In order to be able to run the e2e tests repeatedly on the same cluster, it is necessary to remove the labels on the nodes after every run. This should be taken care of by the test suite.
  • cpmeadors
  • 2
  • Opened 
    17 days ago
  • #448

Pods and instaslice updates are both handled within one reconcile loop. There should be at least 2 reconcile loops for a separation of responsibilities.
  • rphillips
  • Opened 
    on Feb 6
  • #427
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Restrict your search to the title by using the in:title qualifier.
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Restrict your search to the title by using the in:title qualifier.
Issue search results · GitHub