Trim with clustere shared volume and SAN

Hi all,
I have a cluster of 4 Hyper-V 2012 R2 hosts. They are connected to a SAN via fiber channel.
My guest machines range from server 2000 to server 2012 r2.
The LUN provisioned on my SAN (Oracle ZFS3-2 or 7000) are thin provisioned.
My virtual hard disks are both VHD and VHDX.
I need to know how to recover the deleted space from my SAN. Today it just keeps building up until it fills up. I have read about Trim or Unmap but am yet to find a comprehensive doc where I can explore it in detail. I appreciate any help that comes my way.
Try running sdelete inside a VM or shrink VHD(x). First thing will zero out unused blocks and second would do de-allocation on host level. Both should trigger UNMAP on ZFS-based Oracle storage ending with TRIM or whatever. See some links below:
SDelete Utility
http://technet.microsoft.com/en-us/sysinternals/bb897443.aspx
Resize-VHD
http://technet.microsoft.com/en-us/library/hh848535.aspx
Shrinking
VHDX on a running VM
http://blogs.msdn.com/b/virtual_pc_guy/archive/2014/01/30/shrinking-a-vhdx-on-a-running-virtual-machine.aspx
Hope this helped a bit :)
StarWind VSAN [Virtual SAN] clusters Hyper-V without SAS, Fibre Channel, SMB 3.0 or iSCSI, uses Ethernet to mirror internally mounted SATA disks between hosts.

Similar Messages

SQL 2014 FCI+AG on Clustered Shared Volumes design

Hello,
I was assigned the task to design a new SQL 2014 environment for my company. Our current environment consists of two clusters of 4 nodes each (8 nodes total for the two clusters) in two datacenters (let's say LA & NY).
Each datacenter is both primary & DR for different applications. Each cluster is active/active and has 9 SQL instances - 1 DEV/2 UAT/4 PROD/2 DR.
We are using Windows Server 2008 R2 SP1 + SQL 2008 R2 SP2 FCIs for HA + DBM for DR, storage is symmetric, LUNs are coming from NetApp SAN filers.
Due to the consolidated nature of the current SQL clusters (some hold 100+ DBs for different applications), the new platform - Windows 2012 R2 and that one WSFC cluster will be used, I decided to go with a new build instead of upgrade.
My plan for the migration to SQL 2014 is as follows:
1. Evict two nodes from each cluster, the 9 SQLs in each datacenter will run on two nodes and I will have two nodes more per site to build the new environment.
Average CPU utilization for each node is 10-20%, total max memory of all SQL Servers per site is 256GB RAM, each node has 2 CPUs x 8 cores each with 256GB RAM. SQLs will still run on two nodes with reasonable performance.
Since the same number of cores is used and we pay SA, there will be no increase in licensing costs for the time of migration to SQL 2014 and to obsolete the SQL 2008 R2 environment.
2. The new cluster will consists of 4 nodes with Windows 2012 R2 (2 in NY and 2 in LA) and will have Node+FileShare majority quorum model with Cluster Managed voting enabled.
3. New SQL 2014 FCIs will be built (9 per site) and will use Clustered Shared Volumes + AlwaysON Availability Groups. Once the new environment is built, databases will be migrated by backup/restore.
Since we will have AlwaysON FCIs+AlwaysON AGs I'm aware that automatic failover cannot be used and the complications with addition of data files on primary databases, that's fine (file paths will be different of primary and DR SQL instances).
My question is - is this a supported configuration from Microsoft - AlwaysON AGs running on CSVs in a multi-site cluster?
I am aware that CSVs are something new for the SQL world, so what are the things I should take into consideration when building this solution? Are there any pitfalls to be wary of?
Any suggestions will be much appreciated.
Thank you in advance,

First, consider building a separate Hyper-V cluster in each data center, and running SQL Server in VMs.
In your proposed configuration if you loose connectivity between your data centers, half of your nodes will loose quorum and go offline. It's simpler to maintain quorum seperately in each data center when trying to run active-active.
Hyper-V adds a bunch of useful stuff, like live migration, storage migration, and basic HA and DR without guest clustering. And the implementation of Hyper-V on CSV is a very common pattern.
You can then present shared hard disks to your SQL VMs to create a FCI on a guest cluster. See
Deploy a Guest Cluster Using a Shared Virtual Hard Disk
Or you can simply use stand-alone instances on the guests, and use AlwaysOn AGs for rolling upgrades and DR.
That said, you proposed configuration is supported. It's a variant of the solution outlined here
AlwaysOn Architecture Guide: Building a High Availability and Disaster Recovery Solution by Using Failover Cluster Instances and Availability Groups
The cluster would have asymmetric storage, with the nodes in each data center sharing LUNs (whether on CSVs or not), but the LUNs in a data center would not be visible to the nodes in the other datacenter. You can and should use the same path names
on the primary and DR instances, as outlined in the whitepaper. This is important because otherwise file and filegroup operations will suspend the AlwaysOn data movement.
David
David http://blogs.msdn.com/b/dbrowne/

Hyper V hosts with Cluster Shared Volume do not provide Shared VHDX ability

Hi there, i've configured a pair of Hyper V hosts (UCS B230 M2) with SAN-Boot and access to a Shared Fiber Channel LUN.
No matter what I try, when I add a SCSI Adapter and one VHDX virtual disk to a VM, the option for "Advanced" never shows up therefore I can not share the VHDX file between another VM, which prevents any VM based clustering from working (IE: Clustered
SQL server)
Here are the settings I have configured:
The LUN has been formatted with NTFS
The Hyper V hosts have been configured as a Windows Cluster with an IP address and the Fiber Channel LUN configured as a Cluster Shared Volume
I follows the guides below
Create a CSV
https://technet.microsoft.com/en-us/library/dn265980.aspx
Create a shared VHDX
https://technet.microsoft.com/en-us/library/dn282283.aspx
Further details:
Two HyperV server core hosts built
Both joined to AD
Both joined to a Windows failover cluster
One 2TB Fibre Channel LUN presented from an EMC VNX array to both hosts
LUN formatted as NTFS
LUN created as a Cluster Shared Volume
Cluster Shared Volume present on both HyperV hosts
VM created within the Cluster Shared Volume
VM Booted up and Windows 2012 has been installed
VM gracefully shut down
Through the HyperV Manager, going in to VM settings,create a SCSI adapter, create a new VHDX file inside the CSV, the “Advanced” option for the VHDX is not present, therefore the option to make the VHDX file shared is not available
I have followed all steps in the Microsft Technet pages to create a CSV and a VM with Shared VHDX
I’ve cleared all settings and restarted multiple times, but this option is not present

Hi Sir,
>>No matter what I try, when I add a SCSI Adapter and one VHDX virtual disk to a VM, the option for "Advanced" never shows up therefore I can not share the VHDX file between another VM
>>Yes we are using Hyper-V Server 2012 R2
As we know Hyper-v server 2012R2 is a free version without GUI , I'm afraid you are not using Win 8.1 or other windows server 2012R2 to manage that free version hyper-v server .
Shared vhdx is a new feature came with 2012R2 , if use previous windows to manage 2012R2 you will not see "advanced" option (The blue circle belongs 2012 , the red circle belongs to 2012R2 ):
Best Regards,
Elton Ji
Please remember to mark the replies as answers if they help and unmark them if they provide no help. If you have feedback for TechNet Subscriber Support, contact [email protected] .

SQL Server 2014 AG + Clustered Shared Volume

I am testing this scenario in a lab environment:
Node 1 + Node 2 = Failover Instance
Node 3 = Secondary Availability Group.
I am using Shared Clustered Volumes for the Failover Cluster. This causes the volumes to appear on NODE 3 with the same paths.
That is expected.
Problem: If I add a datafile on the failover instance, it tries to add the file to Node 3 using the original path. This fails.
Because, that path already exists and locked as it's Node1/2's cluster volume path.
Any workarounds? With CSV, do I have to have different database locations? Is there any way to modify the clustered volume on one node only or drop from a node in the cluster without dropping the node from the cluster.
Thanks

Hello,
So, I'm not being critical or arguing I'm just responding to your post.
I hope MS does something about that. It is unacceptable for my company (online retail) to suffer through that.
If the resources would have been architected differently, this wouldn't have been an issue. Does it suck that the way you wanted to do it didn't work the way you thought - sure. Are there other ways that get to the same thing? Yep. Were you forced by the
infrastructure teams? Maybe, who know. My point being that it could be done many other ways that wouldn't give you this problem - which to me means it wasn't properly thought out and tested before being implemented.
This is actually called out on page 7/8 of a good white-paper about this:
http://msdn.microsoft.com/en-us/library/jj215886.aspx
In any case, I figured out a hack which should not break anything.
Good luck. Personally, I'd redo the storage.
-Sean
The views, opinions, and posts do not reflect those of my company and are solely my own. No warranty, service, or results are expressed or implied.

Servers with local disk space and SAN disk: is it possible?

Hello,
I am evaluating Oracle VM. We have some servers with local 15000rpm hdds. We also would like to use a SAN. Is it possible to have some domu in local hard disks and some in SAN?
Can I enable HA for domu in SAN?
Thanks in advance for any reply!
Mario Giammarco

I am evaluating Oracle VM. We have some servers with local 15000rpm hdds. We also would like to use a SAN. Is it >possible to have some domu in local hard disks and some in SAN?you can have some domu in local and some in SAN. But domu from local cannot be migrated to other server. In case of SAN you can migrate to other servers.
Can I enable HA for domu in SAN?Yes.

How do I open a folder on my desktop and avoid getting the left hand column with devices shared, etc and the other column with my desktop lists?

I just want to see the contents of the folder not all the other columns!

Finder->View-Hide Sidebar (⌘⌥ S). Once set, ⌘J and select Use as Defaults. That'll ensure all newly opened windows have that setting.

Logical Disk Performance counter for cluster shared volume on Hyper-V

Hello All,
I am trying to collect counters like latency, queuelength from Win32_PerfFormattedData_PerfDisk_LogicalDisk WMI class.
Output of "Name" attribute for logical disks in this class as below:
Name: _Total
Name: C:
Name: E:
Name: HarddiskVolume1
Name: Q:
Name here doesn't show the actual label so I queried Win32_Volume class and wanted to join with performance WMI class. Out put of Win32_Volume is as below:
Caption: E:\
Label: New Volume
Name: E:\
Caption: Q:\
Label: Quorum
Name: Q:\
Caption: C:\
Label: Voume C
Name: C:\
Caption: F:\
Label: SAN
Name: F:\
Please note that "Name" attribute matches for all except one with label "SAN". This is cluster shared volume and "Name" attribute value is "HardDiskVolume1" in Win32_PerfFormattedData_PerfDisk_LogicalDisk class.
Is this is a configuration issue or any other alternative to get volume label and corresponding performance counters.
Thanks in advance
Regards,
Udupa

Hi Udupa,
I haven't found a better way, if you want to combine the two script, please refer to the script below:
$output=@()
$volumes = gwmi Win32_Volume
foreach($volume in $volumes){
$match=($volume.name).TrimEnd("\")
$counter=gwmi Win32_PerfFormattedData_PerfDisk_LogicalDisk |where{$_.name -eq $match}
$Object = New-Object PSObject
$Object | add-member Noteproperty name $volume.name
$Object | add-member Noteproperty label $volume.label
$Object | add-member Noteproperty AvgDiskQueueLength $counter.AvgDiskQueueLength
$output += $Object
$output
I hope this helps.

Color Lables on Shared Volumes

Hi Everyone,
I have just found out that when a user changes the label color for a folder on a shared volume the changes are not reflected real-time. That's to say If I change the colour of a folder called Templates to "Red" another user on another mac who also sees the shared volume will not see the label color change made to the Templates folder. He/She will need to disconnect the shared volume and reconnect in order to see the change. Has anyone out there seen this problem? It's not 100% essential but would be good if there was a solution of some sort.
Just to add did not have this problem on my old mac server which was running 10.2.x our new server is running 10.4.8.
Thanks,
Kwong.

Hi Russell, Sorry been away for a few days. The InDesign documents are already on a shared volume and it wouldn't be feasible to save anywhere else as there is a particular hierarchy on the server as to where files are stored, so that all documents are within their relevant job folders, with images, image bank images & graphics etc stored on other shares. However all acl's are the same for all users, so this shouldn't be an issue. I still can't work out whether this is an Apple OS or an Adobe problem.

Restoring from TC from something other than shared volume

I've got a TC that has several user logins associated with it and I just recently lost a hard drive on my Macbook - trying to restore from one of my TC backups but I'm running into issues.
As suggested by Apple, I didn't place my TC backups in the "Shared" volume of my TC's main drive - What I did do was provide a separate volume under my user login to save them in. Problem is, even when I login as that user through Migration Assistant, it will only bring up the shared volume that the user has access to - not their individual volume that they (and their time machine backups) have access to. Even if I mount my user-specific volume via "Connect to Server", I can see the TC backup but it won't let me choose it in Migration Assistant as it seems to only look at the main shared volume and not any other volume. I am able to mount the TC backup and drag out my files that way but I'd rather actually use the TC like it's meant to be - as a fully-functional restore option. Is there a way to, perhaps by some persuasive UNIX code, to get my mac to restore from that specific volume?

I found a not-so-pretty solution - it requires bringing up Terminal in OSX Installer and following, pretty much, the following directions:
http://www.macosxhints.com/article.php?story=2001120201020569
"sudo" doesn't seem to be recognized in the installer, so you can drop that from the command lines. also, you have to first change directory (cd ..) several times to get to the root folder where you can then do the "mkdir" command correctly. I named mine "sharepoint"
There's no reason to unmount it, you'll want it to stay mounted anyways.
This is nuts that you have to do unix code just to use TC like Apple suggests - there should be a built-in radio button that allows you to connect to other volumes on the TC.

Error -50 on a shared volume (SOLUTION)

Today my antique Xserve (G5, 2 Gigs RAM, 10.4.11) started throwing error -50s on any changes on the main shared volume. Couldn't create a folder, save a file, rename a file or any other change.
After running Disk Utility repairs on the shared volumes and fixing the errors, the volume returned to normal service. YMMV on 10.5 and 10.6 server,but check the disk anyway,

How have you set up sharing.
Can you confirm the exact wording of the error message.

Print authentication problem with a shared USB printer

I'm trying to assist my brother with the setup of his new iMac i3 (running 10.6.4). He is converting from a G5 iMac (which was, I believe, on Tiger) to the i3 and has a Canon MP610 printer. The Canon MP610 printer has been connected via USB to the G5 with printer sharing on, and his wife has been able to print to the printer over Bonjour from her MacBook Pro (OS 10.5.8) when it was shared via the G5.
As soon as we moved the printer to the i3, she became unable to print from the MacBook Pro, because it began putting documents on hold in the queue and requiring authentication to print.
Nothing we've tried will ether allow us to get past the authentication or get rid of the prompt. We've deleted and re-added the printer multiple times, assured that the new i3 is the only one set up for sharing its connected printer - tried different user passwords to authenticate - we're stuck. My brother is the sole user/administrator account on the i3, his wife is the sole administrator/user on the MacBook Pro. The MP610 has been totally deleted from the old G5 so we don't think it is interfering.
They're in California, I'm doing remote support from Texas. The computers and printer are all in the same room in their home.
I've read posts this morning in the community about this issue but they all reference Windows networks.. there is no Windows or mixed Windows-Mac model in our scenario.
Please help! thanks in advance.

The layout of the print window is not the same as OS 9.x and prior. There will not be a choice for double sided printing as there was pre OS X unless the printer comes with an automatic duplexer.
You will have to figure out how to print odds and evens in separate jobs and do it manually.
Matt

When I try to watch a movie on Apple TV from my iPad with home sharing in I only get sound

When I try to watch a movie from my iPad through my Apple TV I only get sound. We have tried it with home sharing on AND off. What am I doing wrong?

Thank you for your feedback

Keep losing connection to back up volumes and TC isn't backing up all data.

Just got this and set it up and having nothing but problems. Wireless network was easy to set up and works reliably (just like my AP Extreme did), however I keep losing a connection with the backup volumes and it is only backing up 95GB of my 106GB of hard drive info. Very hard to tell what's been backed up and what hasn't. Is there anyway of telling what it has grabbed and what it hasn't?
Do I have a bad hard drive? I've given up on using this as a external hard drive and backup, but really bummed on TC. Not at all as easy to use as the AP Store guy said it was...

thanks!

2 Cluster Shared Volumes filling up with shadow copies and I can't delete them.

three node Server 2012 R2 cluster
2 CSVs
Veeam Backup software v7
We've been running out of space on our CSVs. I determined that the hidden System Volume Information directory is huge. It is huge because of large VSS files in the directory.
Path to CSVs
c:\ClusterStorage\Volume1
c:\ClusterStorage\Volume2
===========================================================================
List of VSS providers
vssadmin list providers
vssadmin 1.1 - Volume Shadow Copy Service administrative command-line tool
(C) Copyright 2001-2013 Microsoft Corp.
Provider name: 'SmVssProvider'
Provider type: Hardware
Provider Id: {24e4e8a3-69cb-4370-8b7b-c5276cd49765}
Version: 11.10.0001
Provider name: 'Microsoft CSV Shadow Copy Helper Provider'
Provider type: Software
Provider Id: {26d02d81-6aac-4275-8504-b9c6edc5261d}
Version: 1.0.0.1
Provider name: 'Microsoft CSV Shadow Copy Provider'
Provider type: Software
Provider Id: {400a2ff4-5eb1-44b0-8a05-1fcac0bcf9ff}
Version: 1.0.0.1
Provider name: 'Microsoft File Share Shadow Copy provider'
Provider type: Fileshare
Provider Id: {89300202-3cec-4981-9171-19f59559e0f2}
Version: 1.0.0.1
Provider name: 'Microsoft Software Shadow Copy provider 1.0'
Provider type: System
Provider Id: {b5946137-7b9f-4925-af80-51abd60b20d5}
Version: 1.0.0.7
===========================================================================
Vssadmin list shadowstorage doesn't reference the location of my CSVs...
PS C:\Windows\system32> vssadmin list shadowstorage
vssadmin 1.1 - Volume Shadow Copy Service administrative command-line tool
(C) Copyright 2001-2013 Microsoft Corp.
Shadow Copy Storage association
For volume: (C:)\\?\Volume{6bf32257-eccb-11e3-80b5-806e6f6e6963}\
Shadow Copy Storage volume: (C:)\\?\Volume{6bf32257-eccb-11e3-80b5-806e6f6e6963}\
Used Shadow Copy Storage space: 0 bytes (0%)
Allocated Shadow Copy Storage space: 0 bytes (0%)
Maximum Shadow Copy Storage space: 13.6 GB (10%)
============================================================================
PS C:\clusterstorage\volume1> mountvol
Creates, deletes, or lists a volume mount point.
MOUNTVOL [drive:]path VolumeName
MOUNTVOL [drive:]path /D
MOUNTVOL [drive:]path /L
MOUNTVOL [drive:]path /P
MOUNTVOL /R
MOUNTVOL /N
MOUNTVOL /E
path Specifies the existing NTFS directory where the mount
point will reside.
VolumeName Specifies the volume name that is the target of the mount
point.
/D Removes the volume mount point from the specified directory.
/L Lists the mounted volume name for the specified directory.
/P Removes the volume mount point from the specified directory,
dismounts the volume, and makes the volume not mountable.
You can make the volume mountable again by creating a volume
mount point.
/R Removes volume mount point directories and registry settings
for volumes that are no longer in the system.
/N Disables automatic mounting of new volumes.
/E Re-enables automatic mounting of new volumes.
Possible values for VolumeName along with current mount points are:
\\?\Volume{6bf32256-eccb-11e3-80b5-806e6f6e6963}\
*** NO MOUNT POINTS ***
\\?\Volume{6bf32257-eccb-11e3-80b5-806e6f6e6963}\
C:\
\\?\Volume{bd46bae7-015c-4115-88ca-e6e9067fdea1}\
C:\ClusterStorage\Volume2\
\\?\Volume{caf6938a-5faa-4f3b-8e2b-e27d1498dedf}\
C:\ClusterStorage\Volume1\
\\?\Volume{6bf3225c-eccb-11e3-80b5-806e6f6e6963}\
D:\
====================================================================

When veeam backup call VSS, it create cache files that at the end of backup should be purged. I suppose that in some case this process don't work properly and leave orphaned cache files.
it's possible that when you move CSV on another node these orphaned files will be purged.
from the disk management, right click on CSV, properties, can you see the shadow copies tab?

Moving a Cluster Shared Volume with a VM

Hi,
Is that possible to move a Cluster Shared Volume (dedicated to 1 VM) with its associated VM together if the VM live migrates to another node ?
I try to find a way on Internet but nothing ...
Thank you for the answer :-)

Actually, I have an exact block copy in the secondary site that is being replicated using the storage replication engine.......the CSV in site A is read/Write, and in a "paired" status with the exact block copy in site B. While in "Paired"
status, access to the copy in site B is restricted and cannot be written to. So the Coordinator Node has to be on a node in the same site as the Read/Write Primary CSV in order for the VMs to access the CSV. I have tested this and there is no way
around it, so I am looking for a way to control the Coordinator Node Resource so I can keep it with the VM's CSV when I failover my Geo-cluster to the B site. I have a way to reverse the replication and swap sites during the failover, but I can't make
the coordinator Node resource follow that failover and move with the CSV that has all my clustered VMs on?
I can make the Coordinator Node move to Site B if I Pause and Drain the Roles in Site A, but this is not a viable way to move between sites. so I am looking for a way to move the Coordinator Node resource not the CSV as the link you provide shows.

Trim with clustere shared volume and SAN

Similar Messages

Maybe you are looking for