Chinaunix

标题: ESXi5虚拟机增加第二块盘且扩容磁盘容量 [打印本页]

作者: jkdkungfu 时间: 2012-03-27 19:01
标题: ESXi5虚拟机增加第二块盘且扩容磁盘容量
物理主机ESXI 5，虚拟机WINDOWS 2008，安装虚拟2008分配了2T空间。安装好之后，只分了C、D两个区。现在因为业务数量备份量很大，D盘空间不够。想通过增加第二块盘来解决。现在的问题是如何将第二块盘的容量直接扩到D盘里呢？

ESXI5为啥分配磁盘空间最大只有2TB？

请有过这方面经验的哥们分享一下经验。谢谢

作者: dengbao2001 时间: 2012-03-27 20:00
这位朋友，我正好使用i5版本，我虽没有遇到此问题。下面的文章对你有用

ESX/ESXi 3.x/4.x hosts do not support 2-terabyte LUNs
Details
If you present a 2-terabyte (TB) or larger LUN to one or more ESX/ESXi hosts, you may experience these symptoms:

•You are unable to create a datastore on the LUN
•The ESX/ESXi host does not assign a block device to the LUN
•VMkernel log messages may indicate that the ESX/ESXi host cannot read the capacity of the LUN
•The ESX/ESXi host does not report the correct size on an RDM larger than 2TB
•VirtualCenter or vCenter Server reports incorrect size for the LUN
•When attempting to vMotion a virtual machine using an RDM 2TB or larger from an ESX/ESXi 3.5 to an ESX/ESXi 4.0 host fails. These error messages are found in the vmware.log of the virtual machine:

vmx| Msg_Post: Error vmx| [msg.scsi.esx.diskTooBig] Disks bigger than 2TB - 512B are not supported. vmx| [msg.scsi.esx.virtualDevFailed] Unable to create virtual SCSI device for scsi#:#, '/vmfs/volumes/<datastore>/<vm>/<disk>.vmdk' vmx| [msg.moduletable.powerOnFailed] Module DevicePowerOn power on failed.

•In the VMKernel Logs of the ESX Host you see one or more lines similar to this:

vmkernel: 0:00:41:40.802 cpu1:4177)WARNING: ScsiDeviceIO: 1462: Device 'naa.6001e4f0107c3600110a158c807e60bf': Capacity (numBlocks=0x105e40000) exceeds maximum supported device size of 2TB - 512 Bytes

Note: A cold migration to an ESX 4.0 host followed by a power on attempt has a similar result.

•In vCenter Server 4.0 or later, you see this error:

Error during the configuration of the host: Failed to update disk partition information.

•In the hostd.log file, you see entries similar to:

[F57E1B90 verbose 'HostsvcPlugin' opID=3E815AF8-00000080-6e] Failed to get disk partition information for /vmfs/devices/disks/naa.6001e4f0107c3600110a158c807e60bf
[F5AF5B90 warning 'Partitionsvc' opID=3E815AF8-00000081-d1] Unable to get partition information for /vmfs/devices/disks/naa.6001e4f0107c3600110a158c807e60bf
[F5AF5B90 warning 'Partitionsvc' opID=3E815AF8-00000081-d1] Status : 255
Output:
Error : A bug has been detected in GNU Parted. Refer to the web site of parted http://www.gnu.org/software/parted/parted.html for more informations of what could be useful for bug submitting! Please email a bug report to bug-parted@gnu.org containing at least the version (1.8.1) and the following message: Unable to determine the size of /vmfs/devices/disks/6001e4f0107c3600110a158c807e60bf (Numerical result out of range). Unable to get device /vmfs/devices/disks/6001e4f0107c3600110a158c807e60bf
Solution
Note: As of ESXi 5.0, support for LUNs larger then 2TB is available with both VMFS5 datastores and Physical-mode RDMs. However, virtual disks used in virtual machines and Virtual-mode RDMs are still limited to 2TB - 512 bytes. For more information, see vSphere 5 FAQ: VMFS-5 (2003813).

Cause
The largest LUN size which can be presented to an ESX/ESXi host is 2TB - 512 bytes. This limitation is outlined in the Configuration Maximums guide for the appropriate version of ESX/ESXi.

Understanding the Limitation
VMware ESX and ESXi uses the SCSI-2 standard. As per the SCSI-2 standard, ESX uses a READ_CAPACITY (10-byte) call which returns a 32-bit number to represent the size of the LUN.

Although the call is capable of returning a 32-bit number, only the 0x00000000 to 0xfffffffe values can be used. The value 0xffffffff is reserved.

Based on the standard, the largest valid LUN capacity support is (2^32)-1 sectors. This is the math involved in the calculation:

(2^32) = 4294967296
4294967296 - 1 = 4294967295

If we take the above number and multiply it by 512 bytes per sector we get the size limitation in bytes:

4294967295 x 512 = 2199023255040

The limit of 2TB translates to:

•4294967295 sectors
•2199023255040 bytes
Raw Device Mapping Limit
In the case of physical mode RDMs, the SCSI traffic sent by the guest operating system to do reads or writes is passed through by the VMware ESX storage layer to the underlying storage. This allows some guests to break the 2-terabyte limit, but this is an unsupported configuration because the storage stack of the ESX is not designed to handle such capacity.

Note: When migrating and powering on virtual machines, ESX/ESXi 4.0 enforces the 2TB - 512 bytes limit. ESX 3.5 does not enforce this limit.

Workaround
Although the LUN size is limited to 2TB - 512 bytes, the VMFS and datastore volume size limit is much greater. The maximum volume size is 64TB - 16KB. This can be accomplished by leveraging the maximum extents (32) per volume.

If you require a single datastore that is larger than 2TB - 512 bytes in size, you may want to add extents so that the datastore resides on multiple LUNs of 2 TB - 512 bytes or less.

For information on extents, see the ESX Configuration Guide.

Considerations
Keep these considerations in mind when implementing the workaround:

•Once you add an extent to a datastore, you cannot remove the extent.
•The maximum file size which can reside on a datastore is based on the block size of the filesystem. The maximum block size for VMFS version 3 (used by ESX3 and ESX4) is 8MB, which limits the maximum file size to 2TB - 512 bytes in size. For more information about the maximum file size, see the Configuration Maximums for the appropriate version of ESX.
If you install ESX 3.5 Update 4 on a host that has a local disk larger than 2TB, you may have trouble booting. For more information, see Mounting root fails after installing ESX 3.5 when local storage exceeds the 2 terabyte limit (1012757).

作者: dengbao2001 时间: 2012-03-27 20:02
VMFS-5 Overview
Why switch to VMFS-5?

•VMFS-5 has improved scalability and performance.
•VMFS-5 does not use SCSI-2 Reservations, but uses the ATS VAAI primitives.
•VMFS-5 uses GPT (GUID Partition Table) rather than MBR, which allows for pass-through RDM files greater than 2TB.
•Newly created VMFS-5 datastores use a single block size of 1MB.
•VMFS-5 has support for very small files (<1KB) by storing them in the metadata rather than in the file blocks.
•VMFS-5 uses sub-blocks of 8K rather than 64K, which reduces the space used by small files.
•VMFS-5 uses SCSI_READ16 and SCSI_WRITE16 cmds for I/O (VMFS-3 used SCSI_READ10 and SCSI_WRITE10 cmds for I/O).
What are the limitations for VMFS-5?

•VMFS-5 still limits the number of extents to 32 and the total datastore size to 64TB, but the individual extents are no longer limited to 2TB each. For example, a datastore can be created with a LUN size of 64TB, or a datastore can be created with up to 32 extents up to maximum size of 64TB.
•Only pass-through RDMs can be created with a size >2TB. Non-pass-through RDMs and virtual disk files are still limited to 512B ~ 2TB.
•

assthrough RDMs are supported up to ~60TB in size.
•Both upgraded and newly-created VMFS-5 volumes supported the larger Passthrough RDM size.
Non-passthrough RDMs are still limited to 2TB minus 512B.

作者: jkdkungfu 时间: 2012-03-27 21:07
谢谢楼主的文章~~~~

作者: jkdkungfu 时间: 2012-03-27 21:12
能否将第二块盘空间直接扩容到第一块盘的E盘里呢？

disk.jpg (33.06 KB, 下载次数: 48)

作者: dengbao2001 时间: 2012-03-28 11:24
当然是没问题的。

看这篇文章

http://www.petri.co.il/extend-di ... ows-server-2008.htm

作者: jkdkungfu 时间: 2012-03-28 18:01
本帖最后由 jkdkungfu 于 2012-03-28 18:03 编辑

dengbao2001 发表于 2012-03-28 11:24
当然是没问题的。

看这篇文章

谢谢，按照你给的帖子看了一下。扩展卷选择不了。倒是把新加的两块盘扩展了。

Restore12.png (247.38 KB, 下载次数: 50)

作者: dengbao2001 时间: 2012-03-28 20:11
确保你至少有一个没有分配的分区可用，才可以扩展。

从你截图看，你的磁盘全用光了，所以在无法扩展

作者: jkdkungfu 时间: 2012-03-28 23:11

dengbao2001 发表于 2012-03-28 20:11
确保你至少有一个没有分配的分区可用，才可以扩展。

谢谢你的回复，不过我这边确实扩展不了。因为我新加的两块盘联机初始化后，并没有分区使用。直接在E盘上右键里“扩展卷”灰白的，选择不了。

作者: dengbao2001 时间: 2012-03-29 11:36
测试过，无法再不同磁盘进行扩容的。

你可以讲新的磁盘挂载在E盘的一个空目录，然后使用这个目录

为了测试你这个，我的电脑都挂了，刚弄好

作者: jkdkungfu 时间: 2012-03-29 14:36
呵呵，谢谢楼上哥们。我看了一下资料确实不行。因为2008最大只能识别到2TB。受到了MBR分区表的限制，而GUID分区表则没有这个限制。

作者: deadwind 时间: 2012-03-29 14:44
windows下用>2T的盘

google “动态盘”“GPT”

作者: jkdkungfu 时间: 2012-03-29 16:03
恩，谢谢你。因为ESXI下给虚拟机分配的最大容量是2TB，所以我在操作系统里虚拟机里就是把新加的两个盘转换为动态盘后，再扩展使用的。现在是4T。

欢迎光临 Chinaunix (http://bbs.chinaunix.net/)