vmware: VCdriver creates same hypervisor_hostname for different vcenters with same Cluster name

Bug #1329261 reported by Kanagaraj Manickam on 2014-06-12
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OpenStack Compute (nova)
Matthew Booth

Bug Description

Consider the following scenario, where two nova-compute nodes are configured to different vcenter with same cluster named "Cluster1":

Compute-Node #1:
hostname: devstack1
vmware.cluster_names: Cluster1 << Same cluster name

Compute-Node #2:
hostname: devstack2
vmware.cluster_names: Cluster1 << Same cluster name


 In this scenario, VC driver creates the same value for compute_nodes.hypervisor_hostname as mentioned here.

Compute_node | hypervisor_hostname
#1 | domain-c26(Cluster1)
#2 | domain-c26(Cluster1)

 Because same compute_nodes.hypervisor_hostname is created for more than one clusters, it creates the following issues in the horizon:

1. In admin->Hypervisors, the same CPU, Memory utilization is reported for both the clusters
2. In admin -> Hypervisors-> [Click on Hypervisor #1]->, the details page reports the same list of instance for both the clusters.

Though the issue is seen in the horizon, the actual root cause is because vcdriver creates the same hypervisor_hostname for both the clusters.

so this bug is filed to fix this issue, to create the hypervisor_hostname as follows:
<vmware.host_ip (vmware.cluster_name)>

Ex: (Cluster1)

Changed in nova:
assignee: nobody → Kanagaraj Manickam (kanagaraj-manickam)
Changed in nova:
status: New → In Progress
Tracy Jones (tjones-i) on 2014-06-18
Changed in nova:
importance: Undecided → High
Gary Kotton (garyk) on 2014-08-10
Changed in nova:
milestone: none → juno-3
Thierry Carrez (ttx) on 2014-09-04
Changed in nova:
milestone: juno-3 → juno-rc1
Changed in nova:
assignee: Kanagaraj Manickam (kanagaraj-manickam) → Davanum Srinivas (DIMS) (dims-v)
Changed in nova:
assignee: Davanum Srinivas (DIMS) (dims-v) → Kanagaraj Manickam (kanagaraj-manickam)
Gary Kotton (garyk) on 2014-09-11
Changed in nova:
importance: High → Critical
tags: added: icehouse-backport-potential
Tracy Jones (tjones-i) wrote :

sorry - bugs that affect only a particular hypervisor cannot be critical. I'll add this to the vmwareapi-team and mark critical.

Changed in nova:
importance: Critical → High
Changed in openstack-vmwareapi-team:
importance: Undecided → Critical
status: New → In Progress
Michael Still (mikal) wrote :

Is anyone working this bug?

Michael Still (mikal) on 2014-09-25
Changed in nova:
milestone: juno-rc1 → none

Hi MIchael,

This bug fix is under review Review: https://review.openstack.org/99623

There is one +2, waiting for another one.

Change abandoned by Sean Dague (<email address hidden>) on branch: master
Review: https://review.openstack.org/99623
Reason: This review is > 4 weeks without comment, and failed Jenkins the last time it was checked. We are abandoning this for now. Feel free to reactivate the review by pressing the restore button and leaving a 'recheck' comment to get fresh test results.

Removing "In Progress" status and assignee as change is abandoned.

Changed in nova:
status: In Progress → Confirmed
assignee: Kanagaraj Manickam (kanagaraj-manickam) → nobody

Fix proposed to branch: master
Review: https://review.openstack.org/159481

Changed in nova:
assignee: nobody → Matthew Booth (mbooth-9)
status: Confirmed → In Progress

Reviewed: https://review.openstack.org/159481
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=ce690113c620a3c3a25f8f6d2bfdbaf4e7d2b0d1
Submitter: Jenkins
Branch: master

commit ce690113c620a3c3a25f8f6d2bfdbaf4e7d2b0d1
Author: Kanagaraj Manickam <email address hidden>
Date: Thu Jun 12 16:14:23 2014 +0530

    VMware: Ensure compute_node.hypervisor_hostname is unique

    The VMware driver currently identifies a hypervisor as:

      <mo_id>(<cluster name>)

    Cluster name is not unique, and mo_id is only unique in the scope of a
    single vCenter, so the above can produce duplicate names when managing
    clusters with the same name in multiple vCenters. This is especially
    likely to occur in lab setups.

    This change alters the above format to the following, which uniquely
    identifies a cluster globally:

      <mo_id>.<vCenter uuid>

    Closes-bug: #1329261
    Co-Authored-By: Kanagaraj Manickam <email address hidden>
    Change-Id: I2f3b5d224cc653d0465598de0788116e71d1ca0d

Changed in nova:
status: In Progress → Fix Committed
Thierry Carrez (ttx) on 2015-06-24
Changed in nova:
milestone: none → liberty-1
status: Fix Committed → Fix Released
Thierry Carrez (ttx) on 2015-10-15
Changed in nova:
milestone: liberty-1 → 12.0.0
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers