Bug 1202349

Summary:	virtio-rng required sources change should pre-check hosts in a cluster
Product:	[oVirt] ovirt-engine	Reporter:	Michal Skrivanek <michal.skrivanek>
Component:	General	Assignee:	Martin Betak <mbetak>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Nisim Simsolo <nsimsolo>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	---	CC:	bugs, ecohen, gklein, hannsj_uhl, istein, lsurette, michal.skrivanek, nsimsolo, rbalakri, redhat-bugzilla, yeylon
Target Milestone:	ovirt-3.6.0-rc	Flags:	rule-engine: ovirt-3.6.0+ ylavi: planning_ack+ rule-engine: devel_ack+ istein: testing_ack+
Target Release:	3.6.0
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:	virt
Fixed In Version:	3.6.0-11	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2015-11-04 11:31:33 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	Virt	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	1201508

Description Michal Skrivanek 2015-03-16 12:59:15 UTC

Changing the required virtio-rng sources on an existing cluster is potentially a dangerous thing. If one by mistake require hwrng when no host have it configured properly means that all hosts immediately go Non-Operational.
We should prevent these simple mistakes by one-time check of all the hosts in the cluster and don't proceed (to override this check one can move the offending host to Maintenance...I'd check only for "Up" hosts)

Comment 1 Nisim Simsolo 2015-10-08 11:39:18 UTC

Can you please provide verification scenario?

Comment 2 Michal Skrivanek 2015-10-08 11:45:31 UTC

when the change of parameter would cause host to go non-operational you would now get a warning before proceeding.
One example is you edit cluster and newly require hwrng source, and one or more "Up" hosts in the cluster do not have it available.
Today upon saving Cluster changes those hosts immediately go non-operational, now you will know beforehand:)

Comment 3 Martin Betak 2015-10-08 12:19:45 UTC

@Nisim additionally to the "Up" hosts, the UI should report warning for all hosts in the cluster. This is done so the user would have information about missing capabilities also for hosts in e.g. maintenance mode. 

So the UI warning should list all hosts that fail given check.

Comment 4 Nisim Simsolo 2015-10-14 11:30:13 UTC

Verification build:
rhevm-3.6.0-0.18.el6
host RHEL 7.2 (rng is suuported from 6.6 and above)
sanlock-3.2.4-1.el7.x86_64
qemu-kvm-rhev-2.3.0-28.el7.x86_64
vdsm-4.17.8-1.el7ev.noarch
libvirt-client-1.2.17-5.el7.x86_64

Verification scenario:
1. Have 2 hosts under the same cluscter.
2. Edit cluster and enable /dev/hwrng. click ok.
3. Verify Confirmation pop-up message appears:
 Confirm Cluster Edit
  The following hosts may become non-operational under specified configuration:
  Due to unsupported /dev/hwrng source:

    host_2

  Are you sure you want to continue?
4. click ok.
5. Remove one of the hosts that become unreachable and add it again.
6. Run VM under this host, from VM run the command lsof /dev/random verify host is now supporting hwrng.

Test case for verifying pop-up message created.

Comment 5 Sandro Bonazzola 2015-11-04 11:31:33 UTC

oVirt 3.6.0 has been released on November 4th, 2015 and should fix this issue.
If problems still persist, please open a new BZ and reference this one.