yahoo-eng-team team mailing list archive

[Nikola Đipanov <ndipanov@xxxxxxxxxx>]

This seems to be by design i.e. Scheduler can get out of sync, and we
have the claim-and-retry mechanism in place so request for vm3 would
fail and trigger a reschedule.

** Changed in: nova
       Status: Confirmed => Invalid

-- 
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/1370207

Title:
  race condition between nova scheduler and nova compute

Status in OpenStack Compute (nova):
  Invalid

Bug description:
  This is for nova 2014.1.2.

  Here, nova DB is the shared resources between nova-scheduler and nova-
  compute. Nova-scheduler checks DB to see if hv node can meet the
  provision requirement, nova-compute is the actual process to modify DB
  to reduce the free_ram_mb.

  For example, current available RAM on hv is 56G, with ram_allocation_ration=1.0. Within a minute, 3 vm provision requests are coming to scheduler, each asking for 24G RAM.
   
       t1: scheduler gets a request for vm1, assign vm1 to hv
       t2: scheduler gets a request for vm2, assign vm2 to hv
       t3: vm1 is created, nova-compute updates nova DB with RAM=32G
       t4: scheduler gets a request for vm3, assign vm3 to hv
       t5: vm2 is created, nova-compute updates nova DB with RAM=8G
       t6: vm3 is created, nova-compute updates nova DB with RAM=-16G

  In the end, we have a negative RAM with ram_allocation_ratio=1.0.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1370207/+subscriptions