Bug 2217508

Summary: [DDF] I believe elevator=none (el8+) or noop (el7) is recommended for hyperv workloads.
Product: Red Hat Enterprise Linux 8 Reporter: Direct Docs Feedback <ddf-bot>
Component: DocumentationAssignee: Daniel Vozenilek <davozeni>
Documentation sub component: DDF QA Contact:
Status: ON_QA --- Docs Contact:
Severity: unspecified    
Priority: high CC: davozeni, frival, litian, minlei, rhel-docs, xuli, xxiong, yacao, yuxisun
Version: 8.0Keywords: Documentation
Target Milestone: rcFlags: yuxisun: needinfo? (frival)
Target Release: ---   
Hardware: All   
OS: All   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Direct Docs Feedback 2023-06-26 13:29:55 UTC 
I believe elevator=none (el8+) or noop (el7) is recommended for hyperv workloads.

https://access.redhat.com/solutions/109223

https://learn.microsoft.com/en-us/windows-server/administration/performance-tuning/role/hyper-v-server/linux-virtual-machine-considerations

Reported by: gb-millsjt

https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8/html/deploying_rhel_8_on_microsoft_azure/assembly_deploying-a-rhel-image-as-a-virtual-machine-on-microsoft-azure_cloud-content-azure#annotations:7848af03-24d9-4fc2-8166-17b7c7d5481a
Comment 10 Ming Lei 2023-08-17 13:33:07 UTC 
BTW, on Standard_M128m, one scsi disk(hv_storvsc) is actually HDD., and IOPS can
be improved with deadline, and workload is basically one 128 job fio randread:

fio --filename=/dev/sdb --direct=1 --rw=randread --sync=0 --time_based --bs=4k --size=4096M \
	--numjobs=128 --iodepth 64 --runtime=60 --group_reporting --name=test --thread --ioengine=libaio