yahoo-eng-team team mailing list archive

[Rui Chen <chenrui.momo@xxxxxxxxx>]

Public bug reported:

1. code base

$ git log -1
commit e70ff282f8ed92133942f4d878c2b5f8564f83a8
Merge: 9e0e7a3 222085d
Author: Jenkins <jenkins@xxxxxxxxxxxxxxxxxxxx>
Date:   Mon Sep 14 05:44:39 2015 +0000

    Merge "Invalidate AZ cache when the instance AZ information is
different"

2. Reproduce steps:

* background data: 1000 instances, 10000 faults record in each instance
* list all instance details

Expected result:
* nova-api Host CPU load in a reasonable range, like: 85%

Actual result:
* nova-api Host CPU load persist 100%

NOTE: we collect all faults of each instance in db api (all results
loop), and only return the latest one in InstanceList object(all results
loop), I think we should return the latest fault of each instance in db
api, that can avoid an all results loop in InstanceList object.

** Affects: nova
     Importance: Undecided
     Assignee: Rui Chen (kiwik-chenrui)
         Status: New

** Changed in: nova
     Assignee: (unassigned) => Rui Chen (kiwik-chenrui)

-- 
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/1495400

Title:
  list instance details caused host CPU overload if each one instance
  had lots of instance faults

Status in OpenStack Compute (nova):
  New

Bug description:
  1. code base

  $ git log -1
  commit e70ff282f8ed92133942f4d878c2b5f8564f83a8
  Merge: 9e0e7a3 222085d
  Author: Jenkins <jenkins@xxxxxxxxxxxxxxxxxxxx>
  Date:   Mon Sep 14 05:44:39 2015 +0000

      Merge "Invalidate AZ cache when the instance AZ information is
  different"

  2. Reproduce steps:

  * background data: 1000 instances, 10000 faults record in each instance
  * list all instance details

  Expected result:
  * nova-api Host CPU load in a reasonable range, like: 85%

  Actual result:
  * nova-api Host CPU load persist 100%

  NOTE: we collect all faults of each instance in db api (all results
  loop), and only return the latest one in InstanceList object(all
  results loop), I think we should return the latest fault of each
  instance in db api, that can avoid an all results loop in InstanceList
  object.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1495400/+subscriptions