Bug 1801359

Summary:	ESXi Disk.DiskMaxIOSize considerations for iGW
Product:	[Red Hat Storage] Red Hat Ceph Storage	Reporter:	Heðin <hmoller>
Component:	Documentation	Assignee:	Anjana Suparna Sriram <asriram>
Documentation sub component:	Block Device Guide	QA Contact:	Manohar Murthy <mmurthy>
Status:	NEW ---	Docs Contact:
Severity:	medium
Priority:	unspecified	CC:	asriram, kdreyer, vereddy
Version:	4.1
Target Milestone:	rc
Target Release:	Backlog
Hardware:	All
OS:	Other
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:		Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Heðin 2020-02-10 17:44:59 UTC

Description of problem:
Documentation talking about/explaining block sizes in relation to the iGW, to give a better "picture" of the data flow between iGW Client and RBD Image.
An example could be how changing the ESXi setting "Disk.DiskMaxIOSize"[1] from it's default value of 32767, would affect the iGW performance.


Version-Release number of selected component (if applicable):
Supported versions of ESXi.

[1] https://kb.vmware.com/s/article/1003469

Comment 1 Heðin 2020-02-10 20:08:40 UTC

Loosely affiliated, the following parameters could be added alongside Disk.DiskMaxIOSize:
ISCSI.MaxIoSizeKB 
Misc.APDTimeout

And an explanation on what scenarios can trigger the igw to reject requests, such as too slow IO, due to deep-scrub on a NL deviceclass, sharing LUN export with a much faster class. (happens on 3.x, not sure about 4.x)