VMware Communities : All Content - Submit a User Solution

iSCSI Target for Windows 2008 failover cluster

JanisBalderis — Mon, 28 Sep 2009 08:55:56 GMT

First of all I`m not working for StarWind, I`m using their product for a month. And tell me please, where am I wrong in my posts?

Using KernSafe iSCSI Target for ESX 3.5

Anton Kolomyeytsev — Thu, 03 Sep 2009 07:24:40 GMT

Text was removed by Anton Kolomyeytsev himself

Data recovery software for extracting files from corrupt VMDK images

OfficeRecovery — Tue, 08 Sep 2009 15:50:04 GMT

MediaHeal for Virtual Drives retrieves files from corrupted VMware virtual drive images. VMDK and VHD files and binary disks / partition dumps are supported. The extracted files are saved into a new location chosen by user.

Support for corrupt VMDK snapshot drives has been added recently.

Full description of the utility and a free demo are available at

http://www.officerecovery.com/mediaheal-for-virtual-drives/

OfficeRecovery.com

Converter error "The method 'ConverterValidateTargetStorageParams' does not exist on the object"

ahachenberg — Mon, 20 Jul 2009 05:58:02 GMT

Whoops, it seems my initial searches on this error were flawed. Someone already documented it here --> http://communities.vmware.com/thread/218691

Storage virtualization.

MattZhou — Wed, 10 Jun 2009 10:34:31 GMT

Virtualization plays more and more important role in the life of IT. What could be a heap of physical boxes in the past now turns into just a bunch of files and a virtual infrastructure. VMware occupies 45% of the virtualization market and is a recognized leader in providing virtualization solutions.

Speaking about virtualization in general, we understand that virtually everything related to the IT is possible. One can create virtual hardware, network configuration, a single workstation or a set of virtual workstations, servers and so on. Let alone the central management and low costs of implementation, adoption and maintenance in comparison to physical hardware. Virtualization just changed the way IT goes.

With virtualizing hardware, network and OS I/O operations, we should not forget about storage virtualization which rapidly develops. Virtualization of storage helps achieve location independence by abstracting the physical location of the data.

With storage virtualization, multiple independent storage devices, that may be scattered over a network, appear to be a single monolithic storage device, which can be managed centrally. Usually data storage virtualization goes very closely along with shared storage virtualization. One of the major benefits of abstracting the host or server from the actual storage is the ability to migrate data while maintaining concurrent I/O access.

There are three main implementation approaches :

Host-based

Storage device-based

Network-based

As for technology used in virtual storage virtualization, we can mention iSCSI, FCoE, SMB, VTL and others

Backing up virtual shared storage also becomes easier. This is more convenient and safe at the same time to backup a file than create a copy of a physical SAN.

What other implementation scenarios do you use in your environment? What are the main advantages for your IT environment do you find most important in using virtual storage?

Microsoft Windows 2008 failover cluster with iSCSI

MattZhou — Wed, 10 Jun 2009 11:00:35 GMT

Just my $0.02. StarWind resently announced their spin-off from Rocket Division software. They will now develop iSCSI SAN solutions. (the entire press release: http://www.starwindsoftware.com/news/1)

VMware to MS Virtual Server

MattZhou — Wed, 10 Jun 2009 10:45:54 GMT

In case someone else faces with the same type of issues, StarWind Software Inc. recently released it's free convertor for VMware to Microsoft VM conversion and vice versa:

Learn more at:

http://www.starwindsoftware.com/converter

Copying a VM from MAC to PC

vantosscott — Tue, 09 Jun 2009 18:44:16 GMT

I have a Mac (osx 10.5.7) running VM Fusion (2.0.4).
I have 4 or 5 CentOS VM's running on it. I need to copy one of these VM's to a laptop running windows.
Can someone point me to the manual or let me know the easiest way to copy this.

How to fix disappearing eth0 in Gentoo

isketerol — Fri, 24 Apr 2009 01:45:19 GMT

Hi:

This is a note about how I fixed the disappearing eth0 in Gentoo Linux under VMWare Workstation. It should work for Server and ESXi as well.

The issue is that the module for the network card does not autoload. My inelegant solution was to hack /etc/init.d/net.lo to add these two lines:

modprobe pcnet32

depmod -a

after the last of the opening comments. I restarted /etc/init.d/net.eth0 and eth0 showed up and grabbed a DHCp address.

Hope others will find this useful.

Kevin

setup cannot copy vmscsi.sys

HQNetwork — Wed, 22 Apr 2009 16:01:38 GMT

Hey everyone. After reading through tons of posts and threads regarding driver copy errors, I think I have a solution (at least one that worked perfectly for me).

It looks as though the ones getting the vmscsi.sys error during a Windows XP Professional install, are probably using a re-packaged version of windows. Repackaged meaning an ISO slipstreamed with the latest service pack and/or updates, drivers, etc. This is commonly done with programs such as NLite.

However, you don't need to run any of these programs to fix the driver copy errors. All you need to do is change a value in a file, reinject the file back into the ISO, and use that ISO to install.

As explained by Josh in this thread:

http://communities.vmware.com/message/344803#344803

QUOTE:

...doesn't seem like a vmware problem. seems like your XP PRO has the
OEMPREINSTALL set to YES. is this a custom XP install disk? check the
WINNT.sif file in the i386 folder. Make sure the OEMPREINSTALL=no. I
have built custom install before and if the OEMPREINSTALL is set to yes
it will not copy new drivers into the INSTALL directory.

This worked well for me. I particularly used MagicISO to extract the WINNT.sif to a place I could edit it (desktop or folder), I changed OEMPREINSTALL="Yes" to OEMPREINSTALL="No"

Don't forget to disconnect the vmware floppy image or set it in your vm bios to not read on startup (otherwise you'll get a NTLDR boot error).

Interface Renaming after MAC Address Change (Linux)

Heino Gutschmidt — Tue, 17 Mar 2009 09:33:31 GMT

Hi,

after mac address change (e.g. cloning, storage migration, ...) the network interfaces are renamed (eth0 becomes eth3, eth1 becomes eth4, ...). The network setup does not work any more. Removing or editing the persistent network udev rules (/etc/udev/rules.d) every time is quite hard. It is a better idea to replace mac address matching by bus id matching:

File /etc/udev/rules.d/*-persistent-net.rules (e.g. 70-persistent-net.rules for openSUSE 10.3):

SUBSYSTEM=="net", DRIVERS=="?*", ATTR{address}=="11:22:33:44:55:66", NAME="eth0"

replace ATTR{address}=="11:22:33:44:55:66" by

ID="<BUS_ID>"

To get the bus id use: ethtool -i eth0

The line above would look like:

SUBSYSTEM=="net", DRIVERS=="?*",ID="0000:00:11.0", NAME="eth0"

Heino

sFlow traffic monitoring for virtual switches

Peter6343 — Tue, 10 Mar 2009 01:11:07 GMT

InMon just released free software for monitoring traffic in virtual switches:

http://www.inmon.com/products/virtual-probe/index.php

sFlow is a popular, multi-vendor standard (see sFlow.org) for reporting on network traffic. By exporting sFlow, the software allows virtual switches to be monitored by the same tools used to monitor physical switches.

Please try it out and let us know what you think.

Using iSCSI Cake for ESX 3.5

tankyren — Sat, 17 Jan 2009 05:25:08 GMT

iSCSI Cake is a full featured Windows iSCSI Target. It differentiates from other iSCSI targets in that it supports VMDK file as storage source. This feature enables user to share existing virtual machine disk files or boot from iSCSI to existing operating systems. And whenever there is a problem, the VMDK file can be mounted and diagnosed using VMware's mount tool.

Using iSCSI Cake with VMware ESX Server: http://www.iscsicake.com/iscsi/iscsicakeesx.pdf

Boot Windows 2003 from iSCSI Cake: http://www.iscsicake.com/iscsi/iscsicakedisklessen.pdf

Web site: http://www.iscsicake.com

Linked Clones implementation for VMware VDI environment at UC Santa Barbara, ResNet

lamw — Wed, 17 Dec 2008 16:37:08 GMT

Description:

Motivated by heavy disk costs and slow deployment processes, scripts were developed in-house to make VDI administration easier without incurring any additional costs above that which is necessary. The outcome of these scripts is two tools whose usage will be described in this document.

Our current environment is a small installation of about twenty thinclients (and associated virtual machines) that make up our student accessible lab computers. Total storage consumption is hovering around 80 GB on a single datastore with Linked Clones distributed across 2 ESX hosts. Each VM is configured with 2 vCPUs and 1 GB of memory. At the moment, standard applications like office productivity software is being used on the production VMs. A discussion to extend the application set to include engineering and graphics design oriented software is currently underway thanks to the capabilities of HP RGS.

Update - (1/22/2009)

An issue was found with linked clones that are joined to a Windows domain. Win2k/WinXP/Win2k3 domain members change their computer account passwords every 30 days. When linked clones have been sealed in Step 7, the computer account password generated after joining the domain is saved in the snapshot. This becomes an issue on the 30th day (or 7th day in pre-Win2k machines) when the computer account password is changed and the domain controller receives the new password. Step 8 in the procedure defines that the linked clones are reverted to their state immediately after joining the domain. On the 30th day, this causes the old computer account password of the linked clone to be in effect consequently disallowing any communication between the domain controller and the linked clone.

This problem can be solved in five ways (in order of increasing work and accessibility):

Proactive Solutions:
1) Disable machine account password changes on the domain controller as per instructions from:
http://technet.microsoft.com/en-us/library/cc785826.aspx

This can be limited to just linked clone domain members by applying the necessary group policy security setting to these domain members. This is the solution we employed. Any linked clones that are "Powered On" should be "Rebooted" or "Reverted" and then subsequently "Powered On" after updating the GPO.

2) Disable machine account password changes on the master image (before link cloning) as per instructions from:
http://support.microsoft.com/kb/154501
These instructions apply to Windows 2000 and NT. Use the Windows 2000 instructions on Windows XP. This hasn't been tested. It doesn't seem as though the domain controller will override this setting but it is up to the user to find out.

3) Retire linked clones every month which will probably happen often due to image updates.

4) Delete linked clone snapshots before password update occurs on the 30th day and resnap the clones afterwards.

Reactive Solution:
5) Delete linked clone snapshots, unjoin each linked clone from the domain, rejoin each linked clone to the domain and resnap each linked clone using my-vmware-cmd.sh. This can actually be relatively quick and effortless if scripted properly.

Tools and Hardware:

ghetto-esx-linked-clones.sh
my-vmware-cmd.sh
VMware ESX 3.5u3
VMware Virtual Center 2.5u3
HP SAM (Session Allocation Manager) [VDI Connection broker]
HP RGS Sender/Receiver
HP BL460c
HP EVA4400
HP T5730 thinclients

Implementation:

Step 1.

Create a Master VM image with Windows XP Professional, install VMware Tools and all applications that will be available on the Linked Clone VMs (including SAM registration and RGS Sender services). Ensure that the Master VM is set to automatically obtain an IP address via DHCP and that it is on the correct network that linked clones will be operating on.

Step 2.

Shutdown Master VM image.

Step 3.

Create Linked Clones using ghetto-esx-linked-clones.sh.

Step 4.

Create DHCP reservations on Windows DHCP server from the MAC address file that "ghetto-esx(i)-linked-clones.sh" creates using a custom .vbs DHCP IP reservation script.

Step 5.

Power on all newly created Linked Clones using VIClient if resources allow it.

Step 6.

Once powered on, a custom .vbs script utilizing WMI calls to request that the new Linked Clones join our Active Directory domain is executed from a Windows machine.

Step 7.

Once all VMs have been joined to the domain, the snap function of my-vmware-cmd.sh is run on these VMs. This is required to capture the pristine state of the system right after joining the domain.

Step 8.

At this point, the VMs are ready to be utilized by the available thinclients that have the HP SAM client installed on them. Changes that are made to the Linked Clones by users are discarded because the Linked Clones are refreshed every night using the revert function of my-vmware-cmd.sh that is scheduled with a crontab.

Step 9.

When an image needs to be rebuilt/upgraded, the Master VM image is cloned and worked on while the existing Linked Clones are online. Steps 1 through 7 are followed again on the new Master VM image and new Linked Clones are created with special attention towards disabling these new Linked Clones inside HP SAM.

Step 10.

The old inactive Linked Clones are disabled in HP SAM with old active Linked Clones left alone. The new Linked Clones are then enabled inside HP SAM. The old active Linked Clones will be manually disabled once the user logs out.

Step 11.

The purge function of my-vmware-cmd.sh is then used to destroy the old Linked Clones and all old Linked Clones are removed from HP SAM resources.

Note: With the advent of the new HP RGS licensing model, HP hardware is no longer necessary in a VMware VDI environment utilizing this display protocol.

Backup Virtual Machines in the SAN topology

dconvery — Tue, 20 Jan 2009 13:07:16 GMT

Locosta -

In your case, you have two or three choices:

1. You can install HPDP backup agents in each VM and back up as you would with a physical machine. The good thing is that it is a traditional, easy to understand method. The bad thing is that you need to pay for all of those agents and build a VM and OS before recovery in case of disaster.

2. Use VCB. The good thing is that it is LAN free, if done right. The VCB proxy should be attached to the SAN AND be a media server so that it can write directly to tape. I believe HPDP integrates directly with VCB now. Also, no agents are needed for backups. Disaster recovery is easier than with the traditional approach. The bad thing is that the backup will call vcbMounter.exe. Due to a limitation in Winders cmd.exe, the best throughput you will get on a FullVM backup is about 1GB per minute.. Also, if you do not do FullVM backups, the system stae is not backed up.

3. You could use the EVA scripting facility (SSSU), the VI Toolkit for Windows and powershell to snapshot VMs and then perform a disk based snapshot. Good thing is that you will get better throughput because it is array based. Bad thing is a complicated script to manage. You may be able to do something in RSM to facilitate this as well.

The nice thing is that you are using an HP EVA, HP DataProtector and an HP MSL. Everything will integrate fairly easily no matter what you choose. You don't say how many VMs you have, but you do indicate Windows. VCB is not as robust as I would like, so you will probably need one VCB proxy for every 200-300 VMs, depending on the size of the VMs and at least one per cluster. You will probably find that you will wnd up doing a combination of methods. I usually recommend agent based backups for mail servers,flat file backups for database servers and VCBbackups for msot app, web and file servers. Large file servers can leveragearray based snapshots.

Dave

************************
If the automobile had followed the same development cycle as the computer, a Rolls-Royce would today cost $100, get a million miles per gallon, and explode once a year, killing everyone inside.
Robert X. Cringely, InfoWorld magazine

Maintaining End-to-End Service Levels for VMware Virtual Machines Using VMware DRS and EMC Navisphere QoS

gilmania — Wed, 05 Dec 2007 22:58:19 GMT

*
White Paper: Maintaining End-to-End Service Levels for VMware Virtual Machines Using VMware DRS and EMC Navisphere QoS - Applied Technology

*

http://powerlink.emc.com/km/live1/en_US/Offering_Technical/White_Paper/H4089-maint-end-end-srvc-vmware-virt-mach-vmware-drs-emc-navisphere-qos-wp.pdf?mtcs=ZXZlbnRUeXBlPUttQ2xpY2tDb250ZW50RXZlbnQsZG9jdW1lbnRJZD0wOTAxNDA2NjgwMmNiYjFlLGRvY3VtZW50VHlwZT1wZGYsbmF2ZU5vZGU9MGIwMTQwNjY4MDFiZmYxZg__

This white paper describes tests in which Navisphere QoS Manager and VMware's Distributed Resource Scheduler (DRS) ran in a VMware ESX virtual environment using EMC CLARiiON storage systems. The tests clearly demonstrate that with QoS Manager and DRS multiple applications running in a VMware ESX virtual machine can maintain exceptional mission-critical performance while sharing a CLARiiON storage system.

How To Submit - Community-Supported Hardware

ivanlin — Fri, 02 Nov 2007 09:59:48 GMT

I'm having trouble finding out where to submit community supported hardware configurations.

There's a statement in the link http://www.vmware.com/vmtn/resources/communitysupport/ that says "To submit an entry to this list, register or log in at the VMware Store; be sure and join VMTN when you register. Then create a new entry using the links on the lower left."

I've been looking for the link but without any luck, if anyone have the link please let me know. Thank you.

Security storage with Netapp Decru DataFort solution

tyeo — Fri, 26 Oct 2007 15:07:11 GMT

anyone have comfirm solution paper about Decru DataFort with SAN (FC) storage?

Not Able to remove Snapshots using SH script

jiteshk — Thu, 11 Oct 2007 19:26:29 GMT

Hello All,

I am not able to take the backup through SH script. The script is giving me the error message while removing the snapshots of VM. I am facing the same problem with all the VM's. After running the script, I am getting the error message mentions below:

sh DVM.sh ) = 1esnapshot(ITITEST_Snap "rror executing the command "removesnapshots
Run /usr/bin/vmware-cmd -h to see usage information.

Also note that I am able to remove snapshots manually from command line without using the script. The command which I am using to remove the snapshots is given below:
vmware-cmd /vmfs/volumes/STORAGE_LUN1/ITITEST2/ITITEST2.vmx removesnapshots

Script for taking and removing the snapshots:

#NODE
#TAKING SNAPSHOT
#REMOVING SNAPSHOTS
#
#
#
#
#ITITEST2
vmware-cmd /vmfs/volumes/STORAGELUN_1/ITITEST2/ITITEST2.vmx createsnapshot ITITEST_Snap
vmware-cmd /vmfs/volumes/STORAGELUN_1/ITITEST2/ITITEST2.vmx removesnapshots
#RandD1 (Disk1)
vmware-cmd /home/vmware/RandD1/RandD1.vmx addredo scsi0:0
vmware-cmd /home/vmware/RandD1/RandD1.vmx commit scsi0:0
#RandD1 (Disk2)
vmware-cmd /home/vmware/RandD1/RandD1.vmx addredo scsi0:1
vmware-cmd /home/vmware/RandD1/RandD1.vmx commit scsi0:1

If anyone wants more information, please let me know. Please help me out in solving the problem.

Thanks,
Jitesh Kumar

How-to: Increasing the size of a VM from start to finish

strafford — Fri, 28 Sep 2007 20:38:02 GMT

Not sure if this has been posted somewhere but I was not able to easily find it.

_ *Adding Space to Virtual Machines (example for RHEL VMs)</</p>

1.Increase the size of the VM using vmware-vdiskmanager - in this case I am going from 20Gb to 40Gb:

# vmware-vdiskmanager -X 40Gb vm_name.vmdk

2.Make sure that the VMware Server Console sees the newly added space (in Virtual Machine settings).

3.Partition the added space with parted:

* # parted*
* (parted) print*
* Disk geometry for /dev/sda: 0.000-40960.000 megabytes*
* Disk label type: msdos*
* Minor Start End Type Filesystem Flags*
* 1 0.031 101.975 primary ext3 boot*
* 2 101.975 20473.461 primary lvm*
* (parted) mkpart*
* Partition type? primary/extended? primart*
* Partition type? primary/extended? primary*
* File system type? ext2? ext2*
* Start? 20474.000*
* End? 40960.000*
* (parted) print*
* Disk geometry for /dev/sda: 0.000-40960.000 megabytes*
* Disk label type: msdos*
* Minor Start End Type Filesystem Flags*
* 1 0.031 101.975 primary ext3 boot*
* 2 101.975 20473.461 primary lvm*
* 3 20473.462 40954.768 primary*

4.Create the disk space into a physical volume (PV) using the pvcreate command:

* # pvcreate /dev/sda3*

5.Extend the current logical volume group by adding this new PV:

* # vgextend VolGroup00 /dev/sda3*

6.Now extend the logical volume with the additional space:

* # lvextend -L+20G /dev/VolGroup00/LogVol00*

7.Finally, expand the underlying system to exploit the additional space you just added using ext2online:

* # ext2online /dev/VolGroup00/LogVol00*
* ext2online v1.1.18 - 2001/03/18 for EXT2FS 0.5b*

8.Now your system should be ready to go with the added space, check using the df command:

* # df*
* Filesystem 1K-blocks Used Available Use% Mounted on*
* /dev/mapper/VolGroup00-LogVol00*
* 40607080 18185924 20360308 48% /*
* /dev/sda1 101086 9133 86734 10% /boot*
* none 257852 0 257852 0% /dev/shm*

/usr/sbin/vmkiscsi-ls

Rajesh_interoute — Wed, 26 Sep 2007 08:57:55 GMT

Hi All,

I have iscsi configured on my ESX 3.0.2 server, netapp is acting like iscsi server.

My Netapp has 2 IP address assigned to it , Same LUNs have been made visible to ESX 3.0.2 server via 2 diff IP address ( both IP belonging to same Netapp Storage )

I checked the /usr/sbin/vmkiscsi-ls and saw which IP is Active, as you can see from the output below 10.255.1.10 is the Target Address and 10.130.1.10 is the backup address..

~~root@Server1 root~~# /usr/sbin/vmkiscsi-ls
*******************************************************************************
Cisco iSCSI Driver Version ... 3.4.2 (16-Feb-2004 )
*******************************************************************************
TARGET NAME : iqn.08.com.netapp:aggra.filer.test.internal
TARGET ALIAS :
HOST NO : 0
BUS NO : 0
TARGET ID : 0
TARGET ADDRESS : 10.255.1.10:3260
SESSION STATUS : ESTABLISHED AT Thu Aug 30 11:11:41 2007
NO. OF PORTALS : 2
PORTAL ADDRESS 1 : 10.255.1.10:3260,1002
PORTAL ADDRESS 2 : 10.130.1.10:3260,1001
SESSION ID : ISID 00023d000001 TSID 4d0
*******************************************************************************

I need to get rid of 10.255.1.10 IP address, so now my question is what will happen to the LUNs when i Change that IP on Netapp ?

Will the LUNs be still visible to ESX via 10.130.1.10 IP address ?

Any Commets/Suggestion is welcom..

Regards,

Rajesh

\[solved] Windows NT4 multiprocessor HAL gives 100% CPU load

DoctorNet — Thu, 13 Sep 2007 01:00:39 GMT

Hi all,

After searching on the forum and on google I found, that I am not the only one to have the problem of Windows NT4 mith multiprocessor HAL eating up 1-2x100% CPU in a virtual machine.

Ref:
http://www.vmware.com/community/thread.jspa?messageID=415307
http://www.vmware.com/community/thread.jspa?messageID=418909&#418909
http://www.vmware.com/community/thread.jspa?threadID=37856

While forum user nick.couchman pointed me in a good direction (Thanks Nick if you read this), I couldn't find a HOWTO writeup anywhere. So here we go:

1.) Disclaimer
This worked for me on vmware server 1.01 running on openSUSE 10.2 i586 on a quad-cpu Proliant. It might or might not work for you, it might eat your servers and data - use at your own risk!

2.) The problem and alternative solution
When running Windows NT4 Server (afaik also Workstation) as a guest with the multiprocessor HAL installed (i.e. surely if you assign 2 virtual processors) this virtual machine will allways use 100% CPU time on as many processors as it has assigned.

This is bad, especially since vmware has no chance to figure out, if this CPU-consumption is really needed, or if it is just a guest OS Bug i.e.: An idle virtual NT4 machine will create havoc on the scheduling algorithms, potentially affecting other virtual machines. This interaction between the schedulers of guest, vmware and host really do hurt: see the end of this posting for performance numbers.

Repeat: This problem happens ONLY if you use Windows NT4 SP6a as a guest OS with the multiprocessor HAL installed.

The background seems to be, that the idle process in NT4 SP6a MPHAL does not correctly issue the HLT command to the CPU(s) on idle. This makes the CPU work full power, even when idling - the big amount of heat produced by idle MP NT4 Servers is a hint, that this might really be the issue.

If you have the choice to use a single virtual processor, consider using a singleprocessor HAL (as in http://www.vmware.com/support/ws3/doc/ws32_disks7.html) and get rid of the problem - this is the clean way.

There are some scenarios where the clean way is blocked. Performance is one, in our case we were forced to give 2 virtual processors to NT4 for stupidity reasons of a legacy application: There vendor sold us a license for an application to run on 2 processors some years ago, explicitly allows usage in a virtual environment, but the application will not start on a single processor system for licensing reasons ...

3.) Prerequisites
You will need:
- SrvStart by Nick Rozanski (http://www.nick.rozanski.com/software.htm, GPL-licensed) or another "Run application as service"-tool of your choice

4.) How to

We solve the problem (better: we work around it) by running "CpuIdle" which does in fact nothing but issue HLT to the CPU when it is scheduled, but always remains runnable. So all unused CPU time is soaked up by CpuIdle which has become de facto a new idle process

The HLT goes down to vmware and reduces (heavily) the CPU-consumption.

- Watch CPU usage on Host and Guest: On the guest you will see 100%, most of it for CpuIdle, but on the host you will see CPU consumption by this NT4 instance drop from 100% (200% for 2 virtual processors) to the same level as would be used by W2k.

Until now, this works only while you are logged on interactively, as CpuIdle is an application, not a service. To have it run automatically on startup we need to make it a service: I like SrvStart better than SrvAny - it allows for monitoring of the service, so that when you accidentally open CpuIdle on the console and then end it, it is restarted in the background

- "Install" SrvStart by copying it and its 2 DLLs to a convenient place, I chose C:\Programme\srvstart (C:\Programme is C:\Program Files on a german version of NT4)

- write a control file - here is mine:

#C:\Programme\srvstart\srvstart.ini
debug_out=%TEMP%\myservice.log
[cpuidle]
path=C:\WINNT;C:\WINNT\system32
priority=idle
startup_delay=2
wait_time=2
startup=C:\Programme\CpuIdle\cpuidle.exe
startup_dir=C:\Programme\CpuIdle
shutdown_method=kill
auto_restart=y
restart_interval=30

- register the service:

srvstart.exe install_desktop cpuidle -c C:\Programme\srvstart\srvstart.ini

We need to install the service with desktop interaction, at least as long as it is configured to show the splash screen - this is always during evaluation and after registration as long as you don't explicitly turn it off in the configuration.

The service is created with manual startup mode - so don't forget to set it to automatic.

4.) Performance

A (very) quick and rude benchmark shows the following (remember that the physical host has 4 CPUs), 2 vCPUs are given to the only NT4 virtual machine (processing heavy), 2 vCPUs to a WinXP instance being the only other virtual machine running at the time of the benchmark (IO and processing balanced), RAM is not an issue (6GB present, 1.5 GB for the two running VMs)

- Load average dropped from ca. 1.5 to 0.4

I do not pretend to understand this, but I like it: I can reduce from 3 physical servers to 2 and still survive a desaster.

5.) Further suggestions

I will post to the "user suggestions" of this forum: Would be nice if VMware could include the functionality created by this hack into the VMware Tools

6.) About

Eugen Rieck, Dr. Net! Netzwerkservice, Vienna, Austria
eugen@drnet.at
Programmer and system engineer.

Sorry for the long post!

Best regards, Eugen

This post is in the public domain, but I would appreciate you voluntarily using it in a "Creative Commons Attribution-Share Alike" manner.

ElasticDrive Beta - Remote Amazon S3 Filesystem (VM Data Backup & Recovery)

enomaly — Tue, 31 Jul 2007 21:44:09 GMT

This is a note to inform you that the ElasticDrive S3 Backed FileSystem is now available for free download at http://www.elasticdrive.com/download.html

ElasticDrive is a network block device based upon the Amazon S3 (Simple Storage Service). ElasticDrive provides a caching block device driver which pushes blocks to and from S3 as if they were being written to a local block device.

ElasticDrive provides this service through a virtual NBD service. The NBD service translates from standard NBD to S3 packets transparently, so that the client (the kernel) sees a generic block device. NBD is supported on almost every linux kernel. Simply modprobe nbd, and you are ready to use ElasticDrive.

ElasticDrive is not intended to replace your existing hard drives or network filesystems directly. ElasticDrive is intended to provide seamless backup, RAID target devices, or backing stores for higher level distributed filesystems.

Release notes & Installation instructions
http://www.elasticdrive.com/uploads/media/readme.txt

You can learn more about Enomaly at www.enomaly.net

--

Reuven Cohen
Chief Technologist, Enomaly Inc.
www.enomaly.com :: 212 203 4734 x 1

XP batch script to start necessary services on demand

DAE51D — Tue, 24 Apr 2007 04:56:22 GMT

I've always found it incredibly annoying that VMWare starts four services taking up 12MB of RAM for no reason, when Workstation isn't even running. These just sit there dormant wasting valuable resources (RAM/CPU/startup-time/etc).

What's worse, is that vmware.exe isn't smart enough to start said services if they're not running, when it starts. *sigh*

So... What I do is find the four offending services in XPs Service Manager and set them to start MANUALLY (not Automatic anymore).

Then I create a C:\VMWare\VMWareStart.bat file that contains this
[code]
sc start VMAuthdService
sc start VMnetDHCP
sc start vmount2
sc start "VMware NAT Service"
C:\Utils\Trayconizer.exe C:\VMWare\vmware.exe
[/code]

So then simply change your VMware icon to launch that batch script instead of the vmware.exe and you'll be loving life again.

Oh yeah, it's also very irritating that VMWare doesn't minimize to the tray despite everyone asking for it (http://www.vmware.com/community/search.jspa?q=minimize+to+tray) when you shrink it (or have the option to), so I use this Trayconizer tool to fix that short-coming (ta.

http://www.whitsoftdev.com/trayconizer

Error compiling vmhgfs-only: too few arguments to function get_sb_nodev

mbagni — Fri, 20 Apr 2007 21:14:40 GMT

Inspecting the reason for the error "too few arguments to function 'get_sub_nodev', I found that the reason is the following:

Starting with the release 2.6.18 the function get_sb_nodev declared in fs/super.c has been changed from the previous version 2.6.17.14:

struct super_block *get_sb_nodev(struct file_system_type *fs_type,
int flags, void *data,
int (*fill_super)(struct super_block *, void *, int))

to the following:

int get_sb_nodev(struct file_system_type *fs_type,
int flags, void *data,
int (*fill_super)(struct super_block *, void *, int),
struct vfsmount *mnt)

Since in the file vmhgfs-only calls get_sb_nodev with four arguments, compiling the vmware-tools under kernels newer than 2.6.17.14 is simply not possible, at least with the release 1.0.1-29996 of VMware Server for Linux

Announcement: Linked Clones script for ESX 3.x and other resources

lamw — Wed, 17 Dec 2008 22:29:55 GMT

My colleague (Tuan Duong) and I (William Lam) have been working on a virtualization/VDI deployment project over the last six months. The result of this work is a set of scripts that assist in provisioning and managing the server and lab environment for the Residential Networking Services (ResNet) at the University of California, Santa Barbara.

We took the approach of developing scripts that would be free in nature to support a variety of offerings that currently exist in the enterprise space. One such tool that we would like to share with the VMware community is our Linked Clones script that was developed at the beginning of the summer of 2008. This script functions similarly to the View Composer component in the recent release of VMware View 3 but with relatively relaxed requirements.

A description and more details of the Linked Clones script can be found at:

http://communities.vmware.com/docs/DOC-9020

Another script that complements the Linked Clone’s script is our custom management script “*my-vmware-cmd*” which can be found at:

http://communities.vmware.com/docs/DOC-9061

An example of our implementation of these scripts can be found at:

http://communities.vmware.com/docs/DOC-9201

We also have other scripts and resources that have been consolidated onto a webpage and would like to share it:

http://www.engineering.ucsb.edu/~duonglt/vmware/

We hope that the community finds some of these scripts to be useful in aiding VI administrators to manage their virtual infrastructure and look forward to any feedback that is provided.

Thanks
William lamw and Tuan tlduong

Virtual Appliance - Configuration Analysis & Predictive Fault Management

teich — Tue, 18 Nov 2008 22:57:35 GMT

We've launched a new configuration analysis tool that I hope you guys might find interesting.

Looking at the unified physical and virtual environment, our new product - Replicate Datacenter Analyzer (RDA) - builds up a comprehensive model of a virtualized datacenter through a unique combination of discovery mechanisms. Combining empirical data from Replicate probes with configuration information derived from existing system management tools, RDA constructs a unified view of the datacenter across all administrative domains. RDA analyzes the integrity of the datacenter by leveraging industry best practice and technology dependencies supplied by Replicate in the form of knowledge modules. RDA’s combination of discovery and knowledge driven integrity analysis provides both predictive fault identification and specific resolution guidance for existing and predicted faults.

A few bloggers have had a chance to play with it so far:

We have a free trial version available for download at http://www.replicatetech.com/trial.html

I'd love any feedback or comments - we've built this based on our own pain and our trial users. If you find it useful, we'd love to make it more so!

Oren Teich - Replicate

GUI for Storage Vmotion (svmotion)

robbie800 — Fri, 31 Oct 2008 13:53:30 GMT

Thanks Alex, this tool rocks.

Thanks,

Bob

How to Migrate VMware Server VM's to ESX Server all Disks at once

Vinas — Wed, 22 Oct 2008 15:31:10 GMT

Quick & Dirty: How to Migrate VMware Server VM's to ESX Server using VMs with Multiple Disks
by Josh Davis (Hylant Group, Inc.)

Objective: Move an appliance that has been running on old harware running VMware Server. I want to move this to ESX server 3.x. This is just what I use so if it helps you great! I make no promises so try this on a test system in your environment before taking down production equipment!!!

Audience: Those that have a way to get the VMware Server files to your ESX server. You must be familiar with ssh/linux, and the ESX vmfs file structure.

You will need to be root.

1) Make a folder on your ESX VMFS partition to hold the migrated VM. Copy over the .vmx file, nvram, and everything else but the .vmdk files into this folder.

2) Make a second folder on your VMFS partition (or on the OS root drive perhaps, if it has space), and copy over the .vmdk files.

For single disk based VMs continue with step 3, those with multiple disks see note A below.
3) vmkfstools -i /*path*/*from*/*step2*/disk.vmdk /*path*/*from*/*step1*/disk.vmdk

vmkfstools will convert the media into your new VM directory from step #1. Once verified, you can delete the stuff from step #2.

4) Go into VIC, open up any tab listing your datastores, right-click on the datastore from #1, and pick browse datastore.

5) Drill down to the folder you created.

6) Right-click on the .vmx file and select add to inventory.

7) That should do it!

Don't forget to upgrade the virtual hardware from the pull-down menu, if necessary. Also, you may need to re-install ESX's VMware tools once it's running.

To those using GSX. The virtual LAN adapter sometimes doesn't convert properly. Just delete it out of the VM and re-add it using VIC...

Note A: To do ALL DISKS AT ONCE do this (example 1 is obsolete - use method in example 2):

example 1 (obsolete may not work):
vmkfstools -i /sourcepath/*.vmdk /vmfs/volumes/datastore/hostname/*.vmdk

example 2:
for i in *.vmdk; do vmkfstools -i /*sourcepath*/$i /vmfs/volumes/datastore/hostname/$i; done
(ignore any errors generated by DiskLib_Check, they are generated by the loop)

real world example (my exact syntax, what I used)*:*
for i in *.vmdk; do vmkfstools -i /vmfs/volumes/vm_e1/pba/vmfs/$i /vmfs/volumes/vm_e1/pba/$i; done

That's all I've got!!!

ESX VM backup using Arcserve vmware agent after upgrade to VCB 1.5

_lpl_ — Tue, 21 Oct 2008 08:41:59 GMT

It appears that the Arcserve vmware client agent is no longer able to mount the ESX VM's after upgrading to VCB 1.5.

Also VCBUI.exe (a tool provided by the Arcserve vmware client agent) is not able to mount and add VM's to the Arcserve database.

Nevertheless, VCBmounter runs fine on it's own, so it probably had something to do with communication between Arcserve and VCB.

The content of ca\arcserve backup client agent for windows\checkVCBInstallation.wsf brought me the solution. This file checks if VCB is installed on the local machine (our backupserver)... but when I clicked it, the script told me: "VCB Not Installed".

What do you mean not installed? VCB is located in c:\Program Files\VMWare\VMWare Consolidated Backup Framework... so why do I get this message?

Digging deeper in the script, I found that it looks for a registry key that has a DisplayName = "VMWare Consolidated Backup Framework" below HKLM\Software\Microsoft\Windows\CurrentVersion\Uninstall\

Below this key, the script looks to the InstallLocation value to see where VCB is installed... and this appeared to be empty. Was it made empty after the upgrade to VCB 1.5?

Just filled in the correct installation path of VCB, and our VM backup works perfect again.

Easy End User VM Console Access

rbmadison — Thu, 18 Sep 2008 12:42:09 GMT

I found a bit of aspx code a while ago to allow console access to a VM. It's written in ASP. Attached are the files. Below are the steps you need to do to get this working. Once it's running you simply type in the name or IP of the VM you want to connect to and it brings up the web console of that VM only. If anyone is good at javascript it would be nice to convert this from ASP to javascript so a separate web IIS ASP webserver isn't needed.

1. Create a readonly database user that can access your VirtualCenter database. This is the account that will be used to query the database and get the remote console information.
2. Install IIS on any server other than your VirtualCenter server. You can make this a VM. This is required because we need asp support.
2. Copy the attached getlist.asp and default.asp attached fles into it.
3. Create a system ODBC connection on the IIS server to your VirtualCenter database using the readonly user account.
4. Edit the getlist.asp file and change the following:
a. Change the line "strConnectLog = "Provider=SQLOLEDB..." to reflect your ODBC connection. Use the readonly user account you created in
step 1.
b. Replace all occurances of myvcserver with the host name of your VirtualCenter server.
c. Save the file.
6. Edit the default.asp and replace occurances of myvcserver.mydomain.com with the fully qualified host name of your VirtualCenter server and save
the file.
7. Go to your VirtualCenter server and navigate to \\myvcserver\c$\Documents and Settings\All Users\Application Data\VMware\VMware
VirtualCenter\docRoot. Rename the index.html file to orgindex.html. Don't worry this doesn't break your VirtualCenter and the page will still be
accessible.
8. Save the attached index.html into this docRoot folder.
9. Edit the saved index.html and change the line replace vmconsole.mydomain.com with the fully qualified name of the IIS web site you created in step
2.

Now when you access your VirtualCenter server you will just see a page asking you which console to connect to. Type in a VM IP or host name and click submit.

You can get back to the original default web page by going to the bottom of the page and clicking on the last underscore at the bottom of the page. It will bring up the orgindex.html page. I wanted to make this somewhat hidden for end users but still accessible.

Let me know if I missed anything. If we could get this working in javascript this would be even better so we wouldn't need another IIS server.

Thanks!

Serial Port Mapper

OlgaRitchie — Tue, 26 Aug 2008 12:02:48 GMT

Serial Port Mapper allows to map any serial port to any new port name. In other words, it allows to create aliases for serial ports. E.g.you install additional serial port expansion card to computer as COM 5 serial port. But your favourite software can work with COM1...COM4 ports only and cannot be reconfigured to work with COM5. To resolve this problem you need to map COM 5 port to COM 1 port. As a result your software can open COM 1 port but actually work with COM 5 port. Serial Port Mapper functionality can be easily integrated into your application. So you can map and swap serial ports from your application directly.

How to configure Virtual Manager on ESX 3.5.doc

pareshmd — Mon, 30 Jun 2008 12:54:49 GMT

Rename Datastore

espi3030 — Thu, 19 Jun 2008 13:29:36 GMT

I wanted a way to rename my datastores in my post installation script that finalizes my ESX hosts. With some research I came up with the following:

#!/bin/bash

var1=$HOSTNAME
echo ""
echo "FQDN = $var1"
echo ""

t=${var1//mouse/MOUSE}
u=${t%*...*}

echo "FIXED = $u"
echo ""

ln -sf `readlink -f /vmfs/volumes/storage1` /vmfs/volumes/$u-Local
echo ""

this renamed my datastore from "storage1" to "MOUSE-Local" <-- this is my hostname, which was also renamed from mouse.city.ca.gov. I found out after lots of frustration that a datastore can only be renamed if it is NOT managed by VC server yet.

Add "Notes" to a VM in Virtual Infrastructure Client

espi3030 — Thu, 19 Jun 2008 13:16:43 GMT

Here is a small scritp I came up with to add the backup date in the notes section of a VM:

#!/bin/bash

#SET VARIABLE FOR DATE FORMAT
export date=`date "+%a %b %e %T"`

#GET CURRENT NOTES
for vm in `vmware-cmd -l` ; do
"vmware-cmd" -q $vm getconfig annotation
echo ""

#SET VARIABLE FOR CURRENT NOTES
export ann=`vmware-cmd -q $vm getconfig "annotation"`
ann2=${ann%*Last*}

#ADD NEW NOTES
"vmware-cmd" -q $vm setconfig annotation "$ann2 Last backup done:
$date"
done

With some modifications, I'm sure this script has many other uses. Just thought I'd share this for anyone interested.

Xtravirt Virtual SAN Appliance

oreeh — Wed, 21 May 2008 15:13:35 GMT

Since Alex wants the discussion / questions / comments here
I'm now locking this thread.

Oliver Reeh
VMware Communities User Moderator

Scripts for applying Dell Updates and OpenManage OMSA agent on ESX

ScottHanson — Tue, 20 May 2008 16:01:25 GMT

I've written two scripts to help those with Dell PowerEdge servers in the ESX 3.x environment. The first one will apply the Dell Update Packages (DUPs) through the service console of ESX using the Server Update Utility (SUU).

The details and script are available on our Dell TechCenter site.

The second one automates the manual steps of installing OMSA on ESX, along with configuring the firewall and SNMP settings.

Hope these are a help to some of you.

Scott Hanson - SystemsEdge Blog
Dell Enterprise TechCenter

Enabling and Sharing USB devices over an IP network in a VMware virtual environment

mike@silex — Tue, 29 Apr 2008 22:18:37 GMT

Installing ESX 3.5 on a Sun Ultra-40

publish_or_perish — Tue, 08 Apr 2008 22:44:03 GMT

That's the basic recipe that worked for me. Forgive me if I missed something or if it
doesn't work the same way for you. This should get you close.

Note: The proper 'noapic' setting is critical during install and boot.
noapic is required during installed, but must not be present at regular boots.
If noapic value is wrong, ESX will hang forever during loading of first
module during boot.

The following should be live versions of the links posted in the image above:

+http://www.cs-electronics.com/sas-cables-NEW.htm

SQL Server Reporting Services Snapshot Report Solution

jdsegarra — Mon, 07 Apr 2008 20:25:23 GMT

Tired of answering the question "hey what snapshots are out there on the vm's?" This is a simple report I threw together that queries the Virtual Center database and reports all existing snapshots on the vm's ordered by ESX host. This is dynamic so you can add however many hosts you want and it will automatically pick up the changes. Lemme know what you think, this is my first foray into Reporting Services. Note you'll have to change the data source connection obviously to tailor to your enterprise. I created an account that has read only rights against the VC database for this project. If I'm missing anything in terms of files to upload please let me know. Thanks.

Windows NT 4.0

ialvaradohn — Sat, 05 Apr 2008 20:18:54 GMT

Hi Everybody,

I need some help on a Windows NT 4.0, I've just converted this OS from a compaq Proliant and now that it has been converted to a virtual machine, it seems not to configure, nor recognize any network adapter. I have no knowledge on how to configure a Windows NT 4.0, so does anyone know who to configure the network adapter in it. I've checked everything and nothing seems to work, run the ipconfig command on the command line, and the name of the network adapter shown has the following name = PPP adapter NdisWAN.

Storage VMotion GUI Tool (with option to keep a disk on datastore behind)

azn2kew — Fri, 21 Mar 2008 19:35:09 GMT

I've just downloaded and will testing it right now via my ESX Sandbox to check it out. Thanks for the solution and great done.

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!

Regards,

Stefan Nguyen
iGeek Systems LLC.
VMware, Citrix, Microsoft Consultant

Virtual Appliance - Solve Capacity Bottlenecks (CPU, memory, and storage)

azn2kew — Fri, 21 Mar 2008 19:30:23 GMT

Dave,

I've downloaded and definitely will check it out to make sure its doing what we need. Most of the time, when deploying ESX projects to client, we want to have standard tools to monitor resources even though there are scripts, free tools and commercial we still want to expand our options and hopefully this is going to be good so we're recommend to our clients.

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!

Regards,

Stefan Nguyen
iGeek Systems LLC.
VMware, Citrix, Microsoft Consultant

Capacity Bottleneck Analyzer Virtual Appliance

bradley4681 — Wed, 19 Mar 2008 17:14:14 GMT

I just started using it and I like it so far, helped me find some guests with large disks that didn't really need them. However the "Future Capacity Bottlenecks" graphs haven't been populating with data, so its either not working are I have configured all my systems for unlimited future capacity...

Accessing Virtual Ceneter 2.5/ESX3.5 over VPN Tunnel

mnasir — Tue, 04 Mar 2008 17:49:05 GMT

You will need to do the following to enables access to VC 2.5 using VI client over VPN tunnel:

Open onlty port:443(https) to Virtual Center server

Allow DNS traffic

open port 902(tcp), 903(tcp), to all the esx servers.

This will allow youto access virtual center using VI client and you should also be able to access Virtual machine console. If you olny open port to VC server than you won't be able to access VM console.

Switching to Virtualbox for various reasons - ALSA sound and supported Coherency Mode to name two of them...

Luke Skywalker — Fri, 29 Feb 2008 17:04:14 GMT

You know what REALLY cheeses me? That most important feature that VMware Workstation could offer me is native ALSA support, at least. In the 8-=9 years that I have used VMware, there was always an excuse from someone that Vmware did not want to implement ALSA sound.

Let me put it this way. ALSA is the only real effective and cost prohibitive way to allow applications to share sound from cheap sounds chips (ie: chips that do not allow hardware mixing.) Oh sure, one could implement OSS Multi-app mixing, but why? And OSS is not that stable on my Linux system, especially with the virtual hardware mixing. But ALSA (for you upper-management types - the website is http://www.alsa-project.org) is very stable, very usable, and THE most popular sound system on any Linux distribution. (If you asked Microsoft, they would say, use OSS, so Linux users can stumble in using multiple apps on their their desktop to use sound - right now, Linux and Open Source is the biggest threat to the existance of Microsoft. ).

So since every VMware Linux user for the past 9 years has been waiting for ALSA implementation for a VM Linux host, I announce that I do not have to wait any longer. I am now migrating all my virtual machines to VirtualBox.

(Again for you administrative people, VirtualBox is at - http://virtualbox.org, and it is made by a company that is being bought by Sun Microsystems.

http://blogs.sun.com/tpm/entry/virtualbox

)

Soon, the Java Desktop on Solaris will have integrated Windows install capability. I am sure Sun will buy Windows licenses and include it as a package for the Solaris Java desktop, at least. And maybe they will include Linux too. Anyway, just like the picture of OpenSolaris with Windows Media Player right on the desktop, so I too now have that capability with Linux. So long, lazy over bloated innovation-shackled-by-Microsoft company. Go for the business picture that is BIGGER than Microsoft, not in-bed with Microsoft and limited by Microsoft, trapped in the Microsoft box. And don't talk back to me in your mind, VMware administration. You know what I am talking about. Don't even think of any salesmen shisterisms, let alone saying them. Admit it, that you are weak wussies who can't stand up for the consumer, and a truly free as in freedom, as in a true Democratic, and not Plutocratic) market. The opportunity is there for you as it is for everyone. And if you make any excuses, you are part of the problem, not the solution. There is no excuse for your oversight (not even a monetary one.) There is no try. Only do.

physical to virtual backup

franeborozan — Mon, 18 Feb 2008 23:21:54 GMT

Hello all,

I am more somekind of a lurker on this forums.
Lately I was working on p2v backup project. I will here in short describe what it is about and take a little tame visit the page and tell me what do you think about it. Please note that whole project is open and only free tools are used in implementation.
Whole process is on the www.p2vbackup.com

I will here describe whole process of implementation of virtual servers into your current network infrastructure. With virtualization you will get virtual failover servers and you will have fresh replica of all of your important data, so in case main servers failure you will have fast disaster recovery plan that just works. Of course you can have clusters in your network infrastructure but what is the cost of managing clusters and paying licenses for that expensive kind on in stable software.
There are always savings into hardware and power consumption, imagine now you need to have one physical for every server you own if you want to have good disaster recovery plan, with this you can have dozen of backups on the same server and you will run only backup that is needed in case of the disaster.

big.picture.of project
Main sense of whole project is to keep employees work without interruption caused by server hardware. This process provides low cost failover solution without datacenters or clusters or expensive special hardware and software.
Please note that this project is not best for all server backups, the best would be to contact me with extensive network, servers and data types information and I will suggest is this good solution for you.

Internet connection for VM's

bandhan — Fri, 25 Jan 2008 12:34:02 GMT

window VM IP is 10.50.128.199 and gateway is 10.50.150.1

Backup VM of Windows OS with ImageX

9whirls — Fri, 21 Dec 2007 09:59:37 GMT

Is there anybody interested in this topic?

What's the benefit of doing this?

1. Much smaller disk consumption (compared with zipping each VM into a separate package), because ImageX just stores shared files once if you archieve multiple Windows OS into one WIM file.
2. Easier to share among ESX servers. VM on ESX has a fixed disk size which means it will take much longer time to migrate a vmdk file (equal to the disk size) than doing backup-and-restore via an ImageX image which is based on the actual payload. Furthermore, with ImageX you don't need VC anymore to do VM cloning
3. V2P ? I didn't test this yet, but I believe it's doable, at least logically:). BTW, I believe VMware converter will support ImageX sooner or later. Let's just wait and see.

Software iSCSI netapp filer head failover issues

nmid — Fri, 21 Dec 2007 09:41:43 GMT

i recently implemented a cross site ESX solution using 3.0.2. when performing testing of the storage cluster failover we found that LUNs on the "failed" filer head would not automatically pick up the new filer head and connectivity to storage would be lost

the vmkernel log would display the following error

Dec 15 11:52:30 +esxhostname +vmkernel: 10:19:06:07.450 cpu3:1057)<3>iSCSI:

session 0x7c83a78 portal group tag mismatch, expected 2005, received 2002

performing a manual rescan of the storage adapter would reconnect the storage on fail back of the filer head the ESX host would regain it connectivity to its storage.

Soemthing we found on the netapp support site:

"The Linux, Solaris, and HP-UX iSCSI software initiators create a persistent association between the target portal group tag value and the target.If the target portal group tag changes, the LUNs from the target become unavailable because the initiators recognize the LUNs as new devices."

Both Filers in a clustered pair that use iSCSI require the Portal Tag Group ID to be the same.

This requirement is for Linux and UNIX OSes, since they use the TAG. We have already verified that Windows doesn't seem to use these tags as the iscsi sessions are unaffected by the TAGs.

The Filers create default Portal Groups with TAG ID when iSCSI is enabled on the Filer.

I discovered that these system default TAGs cannot be destroyed or modified.

What we needed to do is create a User-defined Target Group with an ID between 1-256 and assign a Filer interface to it.

In our case we assigned the teamiscsi-1231 interface to the net Group which we called vmware.

An interface can only belong to one Portal Group, so when it is assigned to vmware, it is automatically unassigned from the system default group.

Then you simply repeat the process on the other Filer head

after this change of group id you will need to initially perform a manual rescan fo the storage adapter so it picks up the new ID. after this we found that during a fielr head takeover the ESX server lost minor connectivity during the takeover but automatically regained connectivity when the takeover was complete

we found very little information on this anywhere with the vmware support or forum site so i thought i would post this

Virtual Centre server recovery

pareshmd — Thu, 04 Oct 2007 11:33:28 GMT

Microsoft Windows 2008 failover cluster with iSCSI

mandardhamankar — Thu, 27 Sep 2007 05:26:43 GMT

Aim : To form a 2-node active/passive cluster on Microsoft Windows 2008 guests with iSCSI disks on top of VMware ESX Server.

Procedure : I am assuming that the reader is well versed with Microsoft Windows 2003 Clustering and software-only iSCSI technology. If that is not the case then I would like to request reader, please first understand that and follow the procedure mentioned over here.

1. Three types of disks are supported in Microsoft Windows 2008 failover clustering : serially attached SCSI disks (SAS), iSCSI & FC HBA. We are looking at software-only iSCSI disks inside the guests. In order to achieve this we need to first deploy a VM which will act as a iSCSI target appliance. Configure this VM with multiple virtual SCSI disks (one disk for OS and others to share with guests as iSCSI ove LAN, better to use different SCSI controllers). Install Microsoft Windows 2003 SP1 and then install iSCSI target software (e.g. StarWind iSCSI Target http://www.rocketdivision.com/wind.html). Open iSCSI management console which gets installed as a part of target software and add devices. Add devices is a simple wizard which will create image file out of SCSI disk. This is getting hosted with iSCSI over IP on a default port 3260. Add all the locally attached SCSI disks using add devices wizard. You are done with hosting iSCSI target appliance.

2. To attach, configure and access above iSCSI disks in guest VMs (used for clustering), you need to have iSCSI Initiator software installed and configured. Microsoft provided the same as a part of operating system in their windows 2008. Configure IP or IPs in iSCSI Initiator in both the VMs to talk to above configured iSCSI target appliance. This is also a simple configuration wehre in you are suppose to provide the IP or IPs of iSCSI target applinece and discover the path or paths to disks. Please ensure that the sharing of disks configured correctly and the disks are available even after rebooting of VMs.

3. You can now follow Micorosoft Windows 2008 failover clustering setp-by-step instructions. Validation configuration wizard gives two warnings for Microsoft signed drivers used in VMs against display and mouse. Rest all the tests gets passed successfully which is in-turn a green signal to go ahead and form the cluster.

All the best guys...!

VCBmanager Release 01

skafoelix — Fri, 17 Aug 2007 12:18:20 GMT

VCBmanager is a front-end to vcbMounter and provides a GUI and command line combination to create and delete snapshots.

You can also specify how many concurrent jobs you want to allow,
delay specific hosts,
see which jobs completed successfully or which failed,
view the creation or deletion logs of a job,
see when a job started, when creation started, when creation ended, when deletion started, when deletion ended and when the job finally finished.

Everyone that ever used vcbMounter with any backup software and tried to automate creation and deletion but wanted to keep track of the failed jobs and know why they did fail will appreciate VCBmanager.

http://vcbmanager.sourceforge.net/

VMware Communities : All Content - Submit a User Solution

iSCSI Target for Windows 2008 failover cluster

Using KernSafe iSCSI Target for ESX 3.5

Data recovery software for extracting files from corrupt VMDK images

Converter error "The method 'ConverterValidateTargetStorageParams' does not exist on the object"

Storage virtualization.

Microsoft Windows 2008 failover cluster with iSCSI

VMware to MS Virtual Server

Copying a VM from MAC to PC

How to fix disappearing eth0 in Gentoo

setup cannot copy vmscsi.sys

Interface Renaming after MAC Address Change (Linux)

sFlow traffic monitoring for virtual switches

Using iSCSI Cake for ESX 3.5

Linked Clones implementation for VMware VDI environment at UC Santa Barbara, ResNet

Description:

Tools and Hardware:

Implementation:

Backup Virtual Machines in the SAN topology

Maintaining End-to-End Service Levels for VMware Virtual Machines Using VMware DRS and EMC Navisphere QoS

How To Submit - Community-Supported Hardware

Security storage with Netapp Decru DataFort solution

Not Able to remove Snapshots using SH script

How-to: Increasing the size of a VM from start to finish

/usr/sbin/vmkiscsi-ls

\[solved] Windows NT4 multiprocessor HAL gives 100% CPU load

ElasticDrive Beta - Remote Amazon S3 Filesystem (VM Data Backup & Recovery)

XP batch script to start necessary services on demand

Error compiling vmhgfs-only: too few arguments to function get_sb_nodev

Announcement: Linked Clones script for ESX 3.x and other resources

Virtual Appliance - Configuration Analysis & Predictive Fault Management

GUI for Storage Vmotion (svmotion)

How to Migrate VMware Server VM's to ESX Server *all Disks at once*

ESX VM backup using Arcserve vmware agent after upgrade to VCB 1.5

Easy End User VM Console Access

Serial Port Mapper

How to configure Virtual Manager on ESX 3.5.doc

Rename Datastore

Add "Notes" to a VM in Virtual Infrastructure Client

Xtravirt Virtual SAN Appliance

Scripts for applying Dell Updates and OpenManage OMSA agent on ESX

Enabling and Sharing USB devices over an IP network in a VMware virtual environment

Installing ESX 3.5 on a Sun Ultra-40

SQL Server Reporting Services Snapshot Report Solution

Windows NT 4.0

Storage VMotion GUI Tool (with option to keep a disk on datastore behind)

Virtual Appliance - Solve Capacity Bottlenecks (CPU, memory, and storage)

Capacity Bottleneck Analyzer Virtual Appliance

Accessing Virtual Ceneter 2.5/ESX3.5 over VPN Tunnel

Switching to Virtualbox for various reasons - ALSA sound and supported Coherency Mode to name two of them...

physical to virtual backup

Internet connection for VM's

Backup VM of Windows OS with ImageX

What's the benefit of doing this?

Software iSCSI netapp filer head failover issues

Virtual Centre server recovery

Microsoft Windows 2008 failover cluster with iSCSI

VCBmanager Release 01

How to Migrate VMware Server VM's to ESX Server all Disks at once