yahoo-eng-team team mailing list archive

[Balazs Gibizer <2111440@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

May 21 15:45:11 jammy nova-scheduler[18697]: ERROR nova.scheduler.client.report [None req-45416b86-0807-4fc6-92f4-ce84e7c209d3 demo demo] Failed to retrieve allocation candidates from placement API for filters: RequestGroup(aggregates=[],forbidden_aggregates=set([]),forbidden_traits=set([]),in_tree=None,provider_uuids=[],requester_id='c75d7e16-b840-4b35-b327-7dea81b3736e-0',required_traits=set(['CUSTOM_NVME256G']),resources={CUSTOM_PCI_NONE_NONE=1},use_same_provider=True), RequestGroup(aggregates=[],forbidden_aggregates=set([]),forbidden_traits=set([]),in_tree=None,provider_uuids=[],requester_id=None,required_traits=set([]),resources={DISK_GB=1,MEMORY_MB=256,VCPU=1},use_same_provider=False)
May 21 15:45:11 jammy nova-scheduler[18697]: Got 400: {"errors": [{"status": 400, "title": "Bad Request", "detail": "The server could not comply with the request since it is either malformed or otherwise incorrect.\n\n Invalid resource class in resources parameter: No such resource class CUSTOM_PCI_NONE_NONE.  ", "code": "placement.undefined_code", "request_id": "req-aca3ba2f-cf91-41de-965d-8b37468614c3"}]}.


nova conf:
[pci]
alias = { "name": "nic-foo", "traits": "foo"}


While it is correct that nova fails the scheduling. It should fail with a meaningful log message. 

Also we should think about pre-validating the pci alias config at
service startup instead of only loading and verifying them when a VM is
requesting them.

** Affects: nova
     Importance: Undecided
         Status: New


** Tags: pci pci-in-placement

** Tags added: pci

** Tags added: pci-in-placement

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/2111440

Title:
  PCI in Placement logs an Invalid resource class  error if neither
  resource class nor vendor_id product_id is provided in the PCI alias
  config

Status in OpenStack Compute (nova):
  New

Bug description:
  May 21 15:45:11 jammy nova-scheduler[18697]: ERROR nova.scheduler.client.report [None req-45416b86-0807-4fc6-92f4-ce84e7c209d3 demo demo] Failed to retrieve allocation candidates from placement API for filters: RequestGroup(aggregates=[],forbidden_aggregates=set([]),forbidden_traits=set([]),in_tree=None,provider_uuids=[],requester_id='c75d7e16-b840-4b35-b327-7dea81b3736e-0',required_traits=set(['CUSTOM_NVME256G']),resources={CUSTOM_PCI_NONE_NONE=1},use_same_provider=True), RequestGroup(aggregates=[],forbidden_aggregates=set([]),forbidden_traits=set([]),in_tree=None,provider_uuids=[],requester_id=None,required_traits=set([]),resources={DISK_GB=1,MEMORY_MB=256,VCPU=1},use_same_provider=False)
  May 21 15:45:11 jammy nova-scheduler[18697]: Got 400: {"errors": [{"status": 400, "title": "Bad Request", "detail": "The server could not comply with the request since it is either malformed or otherwise incorrect.\n\n Invalid resource class in resources parameter: No such resource class CUSTOM_PCI_NONE_NONE.  ", "code": "placement.undefined_code", "request_id": "req-aca3ba2f-cf91-41de-965d-8b37468614c3"}]}.

  
  nova conf:
  [pci]
  alias = { "name": "nic-foo", "traits": "foo"}

  
  While it is correct that nova fails the scheduling. It should fail with a meaningful log message. 

  Also we should think about pre-validating the pci alias config at
  service startup instead of only loading and verifying them when a VM
  is requesting them.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/2111440/+subscriptions