[
https://issues.apache.org/jira/browse/CLOUDSTACK-6369?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13964720#comment-13964720
]
Yoshikazu Nojima commented on CLOUDSTACK-6369:
----------------------------------------------
Larger cluster_size accelerates primary storage consumption.
" Smaller cluster sizes can improve the image file size whereas larger cluster
sizes generally provide better performance." - cited from qemu-img man page.
I plan to make it configurable from agent.properties file because this
parameter is specific to KVM/qcow2 and CloudStack admin is an only appropriate
person to configure this parameter.
> Qcow2's cluster_size option support
> -----------------------------------
>
> Key: CLOUDSTACK-6369
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6369
> Project: CloudStack
> Issue Type: Improvement
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Components: KVM
> Reporter: Yoshikazu Nojima
> Assignee: Yoshikazu Nojima
> Priority: Minor
> Fix For: Future
>
>
> To improve the qcow2 performance, it is important to tune parameters of
> "qemu-img create" command.
> preallocation=metadata option, which I am working on in CLOUDSTACK-6191
> improves randam IOPS well, but the qcow2 image created with
> preallocation=metadata option cannot be used as a differential image.
> This "cluster_size" option improves IOPS of a differential qcow2 image which
> is created with preallocation=off.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
