openstack/openstack-manuals
Code Issues Proposed changes

Merge "Added doc for storage policy support in Cinder"

Browse Source
This commit is contained in:
Jenkins 2014-04-26 09:15:51 +00:00 committed by Gerrit Code Review
commit eb17e2680c
1 changed files with 248 additions and 81 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

323
doc/config-reference/block-storage/drivers/vmware-vmdk-driver.xml
View File

@ -6,40 +6,39 @@
<title>VMware VMDK driver</title> <title>VMware VMDK driver</title>
<para>Use the VMware VMDK driver to enable management of the <para>Use the VMware VMDK driver to enable management of the
OpenStack Block Storage volumes on vCenter-managed data OpenStack Block Storage volumes on vCenter-managed data
stores. Volumes are backed by VMDK files on data stores using stores. Volumes are backed by VMDK files on data stores that
any VMware-compatible storage technology such as NFS, iSCSI, use any VMware-compatible storage technology such as NFS,
FiberChannel, and vSAN.</para> iSCSI, FiberChannel, and vSAN.</para>
<warning> <warning>
<para> <para>The VMware ESX VMDK driver is deprecated as of the
The VMware ESX VMDK driver is deprecated as of the Icehouse release Icehouse release and might be removed in Juno or a
and may be removed in Juno or a subsequent release. The VMware subsequent release. The VMware vCenter VMDK driver
vCenter VMDK driver continues to be fully supported. continues to be fully supported.</para>
</para>
</warning> </warning>
<simplesect> <section xml:id="vmdk_functional_context">
<title>Functional context</title> <title>Functional context</title>
<para>The VMware VMDK driver connects to vCenter, through <para>The VMware VMDK driver connects to vCenter, through
which it can dynamically access all the datastores visible which it can dynamically access all the data stores
from the ESX hosts in the managed cluster.</para> visible from the ESX hosts in the managed cluster.</para>
<para>When you create a volume, the VMDK driver creates a VMDK <para>When you create a volume, the VMDK driver creates a VMDK
file on demand. The creation of this VMDK file is file on demand. The VMDK file creation completes only when
completed only when the volume is subsequently attached to the volume is subsequently attached to an instance,
an instance, because the set of datastores visible to the because the set of data stores visible to the instance
instance determines where to place the volume.</para> determines where to place the volume.</para>
<para>The running vSphere VM is then automatically <para>The running vSphere VM is automatically reconfigured to
reconfigured to attach the VMDK file as an extra disk. attach the VMDK file as an extra disk. Once attached, you
Once attached, you can log in to the running vSphere VM to can log in to the running vSphere VM to rescan and
rescan and discover this extra disk.</para> discover this extra disk.</para>
</simplesect> </section>
<simplesect> <section xml:id="vmdk_configuration">
<title>Configuration</title> <title>Configuration</title>
<para>The recommended volume driver for OpenStack Block Storage is <para>The recommended volume driver for OpenStack Block
the VMware vCenter VMDK driver. When you configure the Storage is the VMware vCenter VMDK driver. When you
driver, you must match it with the appropriate OpenStack configure the driver, you must match it with the
Compute driver from VMware and both drivers must point to appropriate OpenStack Compute driver from VMware and both
the same server.</para> drivers must point to the same server.</para>
<para>For example, in the <filename>nova.conf</filename> file, <para>In the <filename>nova.conf</filename> file, use this
use this option to define the Compute driver:</para> option to define the Compute driver:</para>
<programlisting>compute_driver=vmwareapi.VMwareVCDriver</programlisting> <programlisting>compute_driver=vmwareapi.VMwareVCDriver</programlisting>
<para>In the <filename>cinder.conf</filename> file, use this <para>In the <filename>cinder.conf</filename> file, use this
option to define the volume driver:</para> option to define the volume driver:</para>
@ -48,15 +47,16 @@
drivers support for the OpenStack Block Storage drivers support for the OpenStack Block Storage
configuration (<filename>cinder.conf</filename>):</para> configuration (<filename>cinder.conf</filename>):</para>
<xi:include href="../../../common/tables/cinder-vmware.xml"/> <xi:include href="../../../common/tables/cinder-vmware.xml"/>
</simplesect> </section>
<simplesect> <section xml:id="vmdk_disk_type">
<title>VMDK disk type</title> <title>VMDK disk type</title>
<para>The VMware VMDK drivers support the creation of VMDK <para>The VMware VMDK drivers support the creation of VMDK
disk files of type <literal>thin</literal>, <literal>thick</literal>, disk files of type <literal>thin</literal>,
or <literal>eagerZeroedThick</literal>. Use the <literal>thick</literal>, or
<literal>eagerZeroedThick</literal>. Use the
<code>vmware:vmdk_type</code> extra spec key with the <code>vmware:vmdk_type</code> extra spec key with the
appropriate value to specify the VMDK disk file type. appropriate value to specify the VMDK disk file type. The
The following table captures the mapping between the extra following table captures the mapping between the extra
spec entry and the VMDK disk file type:</para> spec entry and the VMDK disk file type:</para>
<table rules="all"> <table rules="all">
<caption>Extra spec entry to VMDK disk file type <caption>Extra spec entry to VMDK disk file type
@ -86,28 +86,25 @@
</tr> </tr>
</tbody> </tbody>
</table> </table>
<para>If no <code>vmdk_type</code> extra spec entry is <para>If you do not specify a <code>vmdk_type</code> extra
specified, the default disk file type is spec entry, the default disk file type is
<literal>thin</literal>.</para> <literal>thin</literal>.</para>
<para>The example below shows how to create a <para>The following example shows how to create a
<code>thick</code> VMDK volume using the appropriate <code>thick</code> VMDK volume by using the
<code>vmdk_type</code>:</para> appropriate <code>vmdk_type</code>:</para>
<screen> <screen><prompt>$</prompt> <userinput>cinder type-create thick_volume</userinput>
<prompt>$</prompt> <userinput>cinder type-create thick_volume</userinput> <prompt>$</prompt> <userinput>cinder type-key thick_volume set vmware:vmdk_type=thick</userinput>
<prompt>$</prompt> <userinput>cinder type-key thick_volume set vmware:vmdk_type=thick</userinput> <prompt>$</prompt> <userinput>cinder create --volume-type thick_volume --display-name volume1 1</userinput></screen>
<prompt>$</prompt> <userinput>cinder create --volume-type thick_volume --display-name volume1 1</userinput> </section>
</screen> <section xml:id="vmdk_clone_type">
</simplesect>
<simplesect>
<title>Clone type</title> <title>Clone type</title>
<para>With the VMware VMDK drivers, you can create a volume <para>With the VMware VMDK drivers, you can create a volume
from another source volume or from a snapshot point. The from another source volume or a snapshot point. The VMware
VMware vCenter VMDK driver supports clone types vCenter VMDK driver supports the <literal>full</literal>
<literal>full</literal> and and <literal>linked/fast</literal> clone types. Use the
<literal>linked/fast</literal>. The clone type is <code>vmware:clone_type</code> extra spec key to
specified using the <code>vmware:clone_type</code> extra specify the clone type. The following table captures the
spec key with the appropriate value. The following table mapping for clone types:</para>
captures the mapping for clone types:</para>
<table rules="all"> <table rules="all">
<caption>Extra spec entry to clone type mapping</caption> <caption>Extra spec entry to clone type mapping</caption>
<thead> <thead>
@ -130,23 +127,109 @@
</tr> </tr>
</tbody> </tbody>
</table> </table>
<para>If not specified, the default clone type is <para>If you do not specify the clone type, the default is
<literal>full</literal>.</para> <literal>full</literal>.</para>
<para>The following is an example of linked cloning from <para>The following example shows linked cloning from another
another source volume:</para> source volume:</para>
<screen> <screen><prompt>$</prompt> <userinput>cinder type-create fast_clone</userinput>
<prompt>$</prompt> <userinput>cinder type-create fast_clone</userinput> <prompt>$</prompt> <userinput>cinder type-key fast_clone set vmware:clone_type=linked</userinput>
<prompt>$</prompt> <userinput>cinder type-key fast_clone set vmware:clone_type=linked</userinput> <prompt>$</prompt> <userinput>cinder create --volume-type fast_clone --source-volid 25743b9d-3605-462b-b9eb-71459fe2bb35 --display-name volume1 1</userinput></screen>
<prompt>$</prompt> <userinput>cinder create --volume-type fast_clone --source-volid 25743b9d-3605-462b-b9eb-71459fe2bb35 --display-name volume1 1</userinput> <note>
</screen> <para>The VMware ESX VMDK driver ignores the extra spec
<para>Note: The VMware ESX VMDK driver ignores the extra spec
entry and always creates a <literal>full</literal> entry and always creates a <literal>full</literal>
clone.</para> clone.</para>
</simplesect> </note>
<simplesect> </section>
<section xml:id="vcenter_storage_policies">
<title>Use vCenter storage policies to specify back-end data
stores</title>
<?dbhtml stop-chunking?>
<para>This section describes how to configure back-end data
stores using storage policies. In vCenter, you can create
one or more storage policies and expose them as a Block
Storage volume-type to a vmdk volume. The storage policies
are exposed to the vmdk driver through the extra spec
property with the
<literal>vmware:storage_profile</literal> key.</para>
<para>For example, assume a storage policy in vCenter named
<literal>gold_policy.</literal> and a Block Storage
volume type named <literal>vol1</literal> with the extra
spec key <literal>vmware:storage_profile</literal> set to
the value <literal>gold_policy</literal>. Any Block
Storage volume creation that uses the
<literal>vol1</literal> volume type places the volume
only in data stores that match the
<literal>gold_policy</literal> storage policy.</para>
<para>The Block Storage back-end configuration for vSphere
data stores is automatically determined based on the
vCenter configuration. If you configure a connection to
connect to vCenter version 5.5 or later in the
<filename>cinder.conf</filename> file, the use of
storage policies to configure back-end data stores is
automatically supported.</para>
<note>
<para>You must configure any data stores that you
configure for the Block Storage service for the
Compute service.</para>
</note>
<procedure>
<title>To configure back-end data stores by using storage
policies</title>
<step>
<para>In vCenter, tag the data stores to be used for
the back end.</para>
<para>OpenStack also supports policies that are
created by using vendor-specific capabilities; for
example vSAN-specific storage policies.</para>
<note>
<para>The tag value serves as the policy. For
details, see <xref linkend="vmware-spbm"
/>.</para>
</note>
</step>
<step>
<para>Set the extra spec key
<literal>vmware:storage_profile</literal> in
the desired Block Storage volume types to the
policy name that you created in the previous
step.</para>
</step>
<step>
<para>Optionally, for the
<parameter>vmware_host_version</parameter>
parameter, enter the version number of your
vSphere platform. For example,
<userinput>5.5</userinput>.</para>
<para>This setting overrides the default location for
the corresponding WSDL file. Among other
scenarios, you can use this setting to prevent
WSDL error messages during the development phase
or to work with a newer version of vCenter.</para>
</step>
<step>
<para>Complete the other vCenter configuration
parameters as appropriate.</para>
</step>
</procedure>
<note>
<para>The following considerations apply to configuring
SPBM for the Block Storage service:</para>
<itemizedlist>
<listitem>
<para>Any volume that is created without an
associated policy (that is to say, without an
associated volume type that specifies
<literal>vmware:storage_profile</literal>
extra spec), there is no policy-based
placement for that volume.</para>
</listitem>
</itemizedlist>
</note>
</section>
<section xml:id="vmdk_supported_operations">
<title>Supported operations</title> <title>Supported operations</title>
<para>The following operations are supported by the VMware <para>The VMware vCenter and ESX VMDK drivers support these
vCenter and ESX VMDK drivers:</para> operations:</para>
<itemizedlist> <itemizedlist>
<listitem> <listitem>
<para>Create volume</para> <para>Create volume</para>
@ -189,17 +272,101 @@
<para>Although the VMware ESX VMDK driver supports these <para>Although the VMware ESX VMDK driver supports these
operations, it has not been extensively tested.</para> operations, it has not been extensively tested.</para>
</note> </note>
</simplesect> </section>
<simplesect>
<section xml:id="vmware-spbm">
<title>Storage policy-based configuration in vCenter</title>
<para>You can configure Storage Policy-Based Management (SPBM)
profiles for vCenter data stores supporting the Compute,
Image Service, and Block Storage components of an OpenStack
implementation.</para>
<para>In a vSphere OpenStack deployment, SPBM enables you to
delegate several data stores for storage, which reduces
the risk of running out of storage space. The policy logic
selects the data store based on accessibility and
available storage space.</para>
</section>
<section xml:id="vmdk_prereqs">
<title>Prerequisites</title>
<itemizedlist>
<listitem>
<para>Determine the data stores to be used by the SPBM
policy.</para>
</listitem>
<listitem>
<para>Determine the tag that identifies the data
stores in the OpenStack component
configuration.</para>
</listitem>
<listitem>
<para>Create separate policies or sets of data stores
for separate OpenStack components.</para>
</listitem>
</itemizedlist>
</section>
<section xml:id="vmdk_storage_policies">
<title>Create storage policies in vCenter</title>
<procedure>
<title>To create storage policies in vCenter</title>
<step>
<para>In vCenter, create the tag that identifies the
data stores:</para>
<substeps>
<step>
<para>From the Home screen, click
<guimenuitem>Tags</guimenuitem>.</para>
</step>
<step>
<para>Specify a name for the tag.</para>
</step>
<step>
<para>Specify a tag category. For example,
<filename>spbm-cinder</filename>.</para>
</step>
</substeps>
</step>
<step>
<para>Apply the tag to the data stores to be used by
the SPBM policy.</para>
<note>
<para>For details about creating tags in vSphere,
see the <link
xlink:href="http://pubs.vmware.com/vsphere-55/index.jsp#com.vmware.vsphere.vcenterhost.doc/GUID-379F40D3-8CD6-449E-89CB-79C4E2683221.html"
>vSphere documentation</link>.</para>
</note>
</step>
<step>
<para>In vCenter, create a tag-based storage policy
that uses one or more tags to identify a set of
data stores.</para>
<note>
<para>You use this tag name and category when you
configure the <filename>*.conf</filename> file
for the OpenStack component. For details about
creating tags in vSphere, see the <link
xlink:href="http://pubs.vmware.com/vsphere-55/index.jsp#com.vmware.vsphere.storage.doc/GUID-89091D59-D844-46B2-94C2-35A3961D23E7.html"
>vSphere documentation</link>.</para>
</note>
</step>
</procedure>
</section>
<section xml:id="data_store_selection">
<title>Data store selection</title> <title>Data store selection</title>
<para>When creating a volume, the driver chooses a data store <para>If storage policy is enabled, the driver initially
which is connected to maximum number of hosts. This is selects all the data stores that match the associated
meant to reduce the number of volume migrations while storage policy.</para>
attaching the volume to instances. The volume needs to be <para>If two or more data stores match the storage policy, the
migrated if the instance's ESX host cannot access the driver chooses a data store that is connected to the
data store containing the volume. In case of ties, the maximum number of hosts.</para>
data store with lowest space utilization is selected, where <para>In case of ties, the driver chooses the data store with
space utilization is defined by the metric lowest space utilization, where space utilization is
<literal>(1 - freespace/totalspace)</literal>.</para> defined by the
</simplesect> <literal>(1-freespace/totalspace)</literal>
metric.</para>
<para>These actions reduce the number of volume migrations
while attaching the volume to instances.</para>
<para>The volume must be migrated if the ESX host for the
instance cannot access the data store that contains the
volume.</para>
</section>
</section> </section>