Bug 720385

Summary:	Report disk latency (read , write & flush ) for each storage device
Product:	Red Hat Enterprise Linux 6	Reporter:	Andrew Cathrow <acathrow>
Component:	vdsm	Assignee:	Igor Lvovsky <ilvovsky>
Status:	CLOSED ERRATA	QA Contact:	Jakub Libosvar <jlibosva>
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	6.2	CC:	abaron, bazulay, cpelland, danken, iheim, jlibosva, lpeer, mprivozn, rbalakri, ykaul
Target Milestone:	rc	Keywords:	SubFeature
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	vdsm-4.9-108	Doc Type:	Enhancement
Doc Text:	No description necessary	Story Points:	---
Clone Of:		Environment:
Last Closed:	2011-12-06 07:31:04 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:	736040, 740918
Bug Blocks:	720379

Description Andrew Cathrow 2011-07-11 14:43:01 UTC

Measure disk latency (read, write and flush) measured in ms for each virtual 
storage
device in QEMU.

Report this through qemu monitor and QMP for each virtual device (I'd  suggest
through query-blockstat)

Use case:

Allow a management system to poll QEMU via libvirt to record the read and write
disk latency. 
This management system could use this information to report on potential
bottlenecks or I/O issues experienced by virtual machines.

Later this information would be used by Red Hat tools as part of an  "SLA"
management application to set thresholds and guaranteed resources for virtual
machines.

We have a very important virtualization systems management vendor lined up to
provide RHEV and KVM support if we can provide this data

Comment 5 Igor Lvovsky 2011-09-05 15:10:44 UTC

The decision is to use virDomainQemuMonitorCommand for RHEL6.2.
We need to wait for python bindings for virDomainQemuMonitorCommand.

Comment 8 Igor Lvovsky 2011-09-14 13:56:13 UTC

http://gerrit.usersys.redhat.com/#change,935

Comment 9 Igor Lvovsky 2011-09-15 10:15:01 UTC

 Andrew,
qemu expose only accumulated disk latency for read, write and flush operations.
So, I don't think that it's what you need to report in RHEV-M.
But, except it qemu can report also the number of requests, so we can calculate the average latency.
What do you think?

Comment 11 Andrew Cathrow 2011-09-20 08:04:29 UTC

(In reply to comment #9)
>  Andrew,
> qemu expose only accumulated disk latency for read, write and flush operations.
> So, I don't think that it's what you need to report in RHEV-M.
> But, except it qemu can report also the number of requests, so we can calculate
> the average latency.
> What do you think?

Yes, that's correct we want the average.

Comment 19 Dan Kenigsberg 2011-10-09 12:02:15 UTC

removing libvirt dependency since we use libvirt_qemu.qemuMonitorCommand hack in rhel6.2.

Comment 21 Jakub Libosvar 2011-10-19 14:32:54 UTC

Verified vdsm-108

Comment 22 Ayal Baron 2011-11-23 07:36:46 UTC

    Technical note added. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    New Contents:
No description necessary

Comment 23 errata-xmlrpc 2011-12-06 07:31:04 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHEA-2011-1782.html