VMware Communities : Unanswered Threads - Enterprise Strategy & Planning

Large Disk Limitations and the 2TB barrier - What problems will I face in the future

Bill.Morton — Thu, 19 Nov 2009 21:59:20 GMT

I want to post a quick question before I go ahead and move one of our production file servers into the Virtual Environment. I want to make sure that I am considering the future limitations of a VM with large disks attached to it so I don't end up in a tight spot later and want any input as to potential limitations that large disks pose for HA, DRS, SRM, vMotion or any other VM aspect.

The file server currently has a 1.5TB 'data' drive that is close to capacity, so it is reasonable to expect that I will need to expand this out to the 2TB limit.

My Storage side is 3x Equallogic iSCSI SAN with Hardware HBA's, 2 ESX hosts. I backup using Vizioncore vRanger 4.1 DPP over the network, though am evaluating moving to VCB + vRanger to offload the backups and speed things up.

My thought would be to make a 2TB LUN on the SAN just for the data drive, then attach a virtual disk that will allow for snapshot growth during backups of 1.75 TB. If additional space is needed in the future, a second LUN with another virtual disk would be attached. The Boot drive and VM config files would live on a different LUN.

Thoughts on this configuration? Better Ideas?

Things I am trying to avoid: LUNs with extents, using iSCSI within the VM. As I understand it these would cause problems with backups.

Q's on setting up new ESX infrastructure and equallogic SAN

huwy — Wed, 18 Nov 2009 16:59:53 GMT

Hi All,

I've recently setup 3 new ESX v4 servers and an equallogic ps4000x SAN. These are connected via 2 dell powerconnect 4524 switches. If you want to critique my setup I have added attachments detailing the layout. I'd be interested if anyone has comments on my setup.

Anyway I'm at the stage where I can start virtualizing servers. What I'm unsure about is the best way to setup ESX/VMs to take advantage of Equallogic's snapshot technology. We are likely to buy the symantec backupexec agent in the future so it makes sense to set them up correctly now.

I can't seem to find any documents on how to set this up so I wondered if anyone had any suggestions? I will be virtualising exchange and SQL servers amongst others.

I anyone has any suggestions or can point me in the right direction I'd be very grateful.

Cheers,

Huwy

Netapp 2020 SAS vs IBM DS3400

gstarr — Tue, 17 Nov 2009 13:13:22 GMT

Hi,

We are in a process to migrate our servers to virtual infrastructure.

One of VAR recomended Netapp 2020 and bundle that Netapp created

is very nice indeed.

Then another VAr said that DS3400 is heaps faster than 2020.

Did someone compare the two?

How much faster is DS3400?

We will be runnning approx. 12 virtual servers, one of them Exchange, and we will

also have Oracle database on the storage.

Please let me know if you have some clues...

Regards,

HA, V-Motion, DR options

Collin09 — Wed, 11 Nov 2009 23:02:06 GMT

1.) I have been testing the ESX 4.0 and now I want to add another VM host server. Before using the HA, Vmotion or DR, does the server hardware configuration has to be the same? Does it make a difference if I am using a different type of hardware.I have 2 server availble but different configuration.

2.) I am inquiring for a SAN storage that will link to VM host server for disaster recovery. I already have 2 server I can use but trying to find a decent SAN storage. Any recommendation of SAN storage that can hold 2-3TB and a decend price range?

Your input is greatly appreciated.

Thanks.

Collin

Training Documents for ESX4 & VC4?

upg3 — Mon, 09 Nov 2009 14:16:21 GMT

I've got to give a little training session on ESX4 and Virtual Center4 for some folks who are going to manage their own little ESX4/VC4 environment. Does anyone have any powerpoints or electronic training guides they'd like to share with me?

Or any recommendations other than just going through Virtual Center on the big screen?

Thanks!

performance wise setup for SQL?

mlubinski — Fri, 06 Nov 2009 16:14:12 GMT

Hi I would like to ask people who are/were working with SQL servers that should run with optimal performance all the time.

Let's assume, that we want to build up 6 SQL servers, each with 4vCPU+16GB of memory. I have 3 hosts (each host with 2xQuadCore+32GB) plus NFS storage on Netapp (dedicated volume for these servers)

So tell me, if I put 2 such SQL VMs on each ESX host, to make these servers run with no performance issues, should I set them any memory reservations (or maybe they shouldn't have any reservations set)?

Should they have any specific swapfile "location" for them? (in VM configuration->swapfile location). What about internal OS setup? Should it have some OS (Windows 2008 x64) tweaks inside for best performance?

I think this setup should be enough for this, but was wondering what else should I think of. These reservations make me crazy, coz I don't know what in this situation would be the best option.

I hope someone will be able to give me some tips on this.

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points

VSphere Capacity Planning

vmproteau — Mon, 02 Nov 2009 21:14:28 GMT

Ignoring CapacityIQ which I understand isn't yet supported in VSphere currently, I'm trying to develop a logical calculation to estimate capacity in a particular environment. We use an N+1 sizing strategy so, for 2-Quad Core, 32GB memory Hosts, we would max our N+1 memory ceiling and buy another Host before CPU would ever be a concern.

We are now looking at the same amount of cores (2-Quad Core) but increasing to 96GB of memory (HP DL380 G6) so now I'm more concerned with CPU performance. I have heard of some generic VM/Core best practices (4-6 for ESX 3.5) (8-10 for ESX 4.0) however, I'm looking for a more accurate calculation.

Can anyone tell me what metrics I would need to look at and perhaps what calculation I would make to best estimate what kind of CPU capacity I safely have in a paticular Host/Cluster. I'm not even sure if I need to be looking at average MHz over time, vCPU/Core, VM/Core, etc.?

I know there are tools for getting this information like CapacityIQ but, I am trying to "right size" our future Hosts hardware standard.

vMotion Porgroup & Serice Console PortGroup can have the same IP ?

shatztal — Sun, 01 Nov 2009 10:46:37 GMT

hello, in our enivroment the configuration is that the vMotion and Serviceconsole IP are the same on each esx.

the vMotion worked but somehow it didn't and now i am asking if maybe the problems can accure because of his config altough it worked before

also in the /etc/host file i insert all the IP addresses of the ESX in the Clusters.

need some help

Win2003 Multiprocessor to Uniprocessor HAL

NYSDHCR — Fri, 30 Oct 2009 13:59:04 GMT

Is it possible to change from a Win2003 Multiprocessor to Uniprocessor HAL, if the vm was initially built with 2 VCPUs? I tried to remove one vcpu and downgrade the uniprocessor, but it was not an option.

HP DL380 G6: ESX Host Memory or CPU

vmproteau — Sat, 24 Oct 2009 00:05:50 GMT

We have a primarily HP shop where our ESX Hosts are HP DL380 G5, Quad Core, 32GB RAM. Generally memory is the deciding factor on when to consider adding Hosts. CPU isn't really considered because our VM/Core ratio is relatively small.With VSphere and lease replacements looming we are likely looking at HP DL380 G6 Quad Core, 64-72GB RAM. Those familiar with the G6 may know that memory population can greatly affect speed where 1333 MHz memory can be reduced to 800MHz speed depending on configuration.

To maximum memoy speed at the proposed capcities, we'll have to spend 10X on 8GB DIMMs. If we did that we'd probably look at less of a processor. I am concerned with the larger boxes because our VM/core number will certainly increase so my gut tells me to get the fastest processors I can afford but, I have no evidence to back that decision.

If I had to sacrafice one for the other should I go with faster CPU or Memory speeds. I do realize the VM population would factor in. We have a shared environment so server role runs the gamit but, they are generally Windows 2003/2008. Is there any guidance or evidence to help determine if the benefits of faster memory outweigh CPU speed concerns or vice versa?

On a related note, anyone know why the HP DL300 series G6 Intel offerings only Quad Core while the AMD offers Hex. Are Intel Hex in the works?

Periodic clean up of inactive VMs

crowhurst — Fri, 23 Oct 2009 12:58:49 GMT

Can someone point me in the direction of a canned report to identify VMs that are suspiciously inactive? I am thinking something along the lines of very low levels of cpu and disk activity for a prolonged period. I am trying to periodically spot those VMs that are basically just an idle OS and to ping their owners to see if they are still needed. Thanks

DHCP IP-Adress for the ESX Service Console a good plan ?

meistermn — Fri, 23 Oct 2009 09:30:28 GMT

In our ESX Server environment we use today static ip-address for the service console and the vmotion.

Now what will happen, if the service console gets a dhcp address.

Is there a risk.
What does this mean to HA isolation mode ?

Best approach for a Windows WebFarm

alnapp — Thu, 22 Oct 2009 10:55:23 GMT

Hello all

I am going to implement a webfarm on our vmware infrastructure

This will be app'x 50 nearly identical windows machines (2008 32bit)

We are on ESX 3.5 and 4 (mid move)

What clever vmware tricks (I am thinking, consolidated disks etc) could I use in this scenario?

Would vmware VDI be suitable, or is that notapt for nearly identical servers?

Any pointers would be appreciated

Cheers

Windows Storage Server 2008 in ESXi

wibni — Wed, 21 Oct 2009 07:19:31 GMT

Hi,

I've got my first virtualisation project coming up. It's very basic and I wanne keep it that way till I understand more about it.
We're about to implement a new engineering application which I wanne run in VMWare.

I'm about to purchase VMWare vSphere4 Standard and a DELL R710 Server.
It will just be 1 node and we don't have an external storage system so I was thinking of running Windows Storage Server 2008 in 1 VM and the apllication server and database server in 2 more VMs on W2k8 R2 Standard.
So it will be 3 VMS on this ESXi.

It appears that MS does not sell WSS 2008 as a stand alone license - only as OEM together with hardware.
Would anyone here have come accross the same problem already?
I found this article from MS explaining it for WSS 2003. http://support.microsoft.com/kb/942868

Has anyone got Windows Storage Server 2008 running on ESXi and can share whether it causes issues as described in the article from MS?
What would be the better alternative? Just run the file server on W2k8 R2 Standard as well or overthink my design?

CA Spectrum for virtualization anyone?

azn2kew — Tue, 20 Oct 2009 16:14:48 GMT

Does anyone using CA Spectrum in their environment and how well does it work? Anyone know where to acquire trial demo to see how it works? I've tried google and CA site didn't find it. Please provide your feedback why you like it?

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

Regards,

Stefan Nguyen
VMware vExpert 2009
iGeek Systems Inc.
VMware, Citrix, Microsoft Consultant

VMFS Extents for Performance.

RussellCorey — Mon, 19 Oct 2009 20:57:38 GMT

Doing some reading:

http://virtualgeek.typepad.com/virtual_geek/2009/03/vmfs-best-practices-and-counter-fud.html

and of course Scott Lowe's book "Mastering vSphere" :

"Data is distributed across the LUNs within the spanned VMFS volulme."

I would like to understand how this data is distributed.

Presume we have 4 LUNs that are 100GB each on a dedicated front end array port and I create a new a single .vmdk file thats say 40GB.

Is all that 40GB on the first LUN of the extent or is it distribtued on all 4 LUNs?(i.e. 10GB on LUN0, 10 on LUN1, etc.)

If the former, will a new VM be placed on the next LUN?

Can you provide some supporting documentation on whatever behavior?

CPU Reservation Question

heybuzzz — Mon, 19 Oct 2009 19:44:23 GMT

My ESX 3.5 host has 2, quad core 3.00Ghz CPU 's and my host sees that as 8 x 2.99 GHZ. How come when I set a VM to two vCPU's and 6000 MHZ CPU reservations do I get an error that says "Failed to find host for powering on VM"? If I take the reservation away and set it back to zero I can power on the VM fine. If I then set the reservation back to 6000 when the VM is powered on it changes the hosts HA status to disable.... There are no other VM's running on this host when I get the errors... I know there is an equation for this but I haven't done it in a long time. What am I missing here? Thanks

3par F400 and their software functionality

dkaiserlt — Mon, 19 Oct 2009 15:26:44 GMT

Hi,

We are leaning towards 3par's F400 series with dual controllers (active/active) and wanted to find out if anyone has any information to share regarding their experience with the unit and especially how they leverage 3par's software suite (3Par Virtual Copy, 3Par Remote Copy, 3Par Dynamic Optimization) in their vSphere environment.

Thanks.

VMWare Config Advice

paulkc — Mon, 19 Oct 2009 03:02:46 GMT

I am currently running VMWare Server 2 on an Ubuntu Desktop host. Here is the basic configuration of my host:
Consumer grade Gigabyte motherboard
AMD Phenom quad-core processor
8gb ram
Multiple Sata drives

I am running the following VMs on this host
VM1: W2K3 running as DC for the network, also performing DNS and file sharing
VM2: W2K3 running a SQL express db and also a couple of small Access/Paradox based dbs accessed by only 2 or 3 users. Also running an Openfire chat server for roughly 12 users.
VM3: Linux Based (Red Hat) Content Filtering System for roughly 20-30 simultanous users.

I am having trouble with the host freezing up every few days. It seems that the longer it runs, the more trouble I have with the network locking down and things beginning to not respond properly. I am not sure if I am dealing more with a software issue or more of a hardware issue. I would like to move back to a Windows Host as I am not very knowledgeable in anything Linux (which is why my host is Ubuntu Desktop instead of Ubuntu Server).

I would like to actually host my VM's off of an Openfiler NAS and run them off a Windows Based host but I am having so much trouble out of the locally hosted VM's that I didn't want to make that move until I do some more planning/research.

Any thoughts/advice would be greatly appreciated at this point.

Infortrend S16F-R1840 and excessive LUN locking

erickmiller — Wed, 14 Oct 2009 08:22:18 GMT

Has anyone used Infortrend's S16F-R1840 fiber channel SAN?

We've had various problems from day one, which a recent firmware update solved many problems, but we often get SCSI I/O Reservation Conflicts that result in a locked LUN (a LUN inaccessible by all but one host).

I'm quite familiar with the problems that result in re-tried I/O reservations due to conflicts during a reservation. We have some HP and IBM DS3400 SANs that have "no" problems whatsoever, so this is a relatively new problem with a new SAN.

I've been in almost constant communication with Infortrend's support regarding this issue, and thought that it was related to a number of possible problems, but none
have panned out (firmware updates, ESX upgrades to the latest version, BIOS updates, fiber channel switch port error counter checks, etc.).

We have both ESX 3.5 and ESX 4.0 clusters attached. It appears the problem occurs at random times. We've even had it happen on a 2-node cluster by itself. I thought it was related to high I/O, but I've seen a LUN get locked with little I/O. Recently, it occurred while backups were running (using esXpress) while some Storage VMotions were running, so I can understand there will be some conflicts in this case, but nothing that would bring a whole LUN down. Since we
push our HP and IBM SANs to their limits with the same workloads, and have no problems at all, I'm confident in saying that this is an Infortrend problem.

All hardware is on the HCL (including the Infortrend), and includes mostly HP DL385 G5p's with FC2142SR (Emulex LPe1150) controllers or HP DL365 G5's with FC2242 (Emulex LPe11002) controllers (dual-port version of the LPe1150). Fiber channel switches are HP StorageWorks 4/16 switches (Brocade Silkworm 200E switches).

So, I'm looking for others that have had similar issues since this driving me nuts and we're about ready to return this unit (not going to be a great day if it comes to that).

Eric K. Miller, Genesis Hosting Solutions, LLC
http://www.genesishosting.com/ - Lease part of our ESX cluster!

ESX and BizTalk 2009 R2

Iberia VMware User Group — Tue, 13 Oct 2009 14:31:56 GMT

HI, does anyone know a paper, link, document, or anything about wich version of ESX is compatible with BizTalk 2009 R2?

Tanks,

Ressource pools and minimum reservation

idaho51 — Tue, 13 Oct 2009 08:08:59 GMT

Dear all,

since a couple of years we are using VMWare, but currently we have an internal debate on the correct behaviour of ressource pools and minimum ressource assignment to either the pool and/or single VM's.

As far as documentatino goes the ressource pools should only come into to play when a cluster/physical host become congested for cpu or memory ressources, and then apply the configured shares to the ressource the VM's are asking for.

So can somebody shed some light whether ressource pools are actually used in non congested environments?

Second the usage of mininum ressource assignment. Minimum ressources can be assigned to a VM directly or the the upper layer ressource pool, logically the assignend ressource should then be downhanded to the VM's in this ressoruce pool, is this correct?

Does this the assignemt of minimum ressources also works between a parent ressource pool and VM' which are in a child ressoruce pool? And last does the assignment of minimum ressources (CPU and Memory, e.g. 256MHz and 256MB Ram) improve the performance of VM's a kind of kick start to the VM in case it needs ressources.

Thank you all for your contribution in advance.

Best regards,

Andreas

How to become a VMWARE consultant or solutions architect or VMWARE Architecture

venkat82 — Mon, 12 Oct 2009 11:11:06 GMT

Hi ALL,

I have completed my VCP 3.5 exam and i am interested to become a vmware consultant or architecture ....Right now i have 6 years exp and has exp in vmware support. Please advice for further my career path, how do i go aout it.. please advice....

Regards

Venkat

Hardware Suggestions

pkuzma — Sun, 11 Oct 2009 17:37:53 GMT

I am looking to utilize puppylinux and rdesktop on current and future desktops to centralize about 90 workstations. I plan on running 90 seperate virtual machines as each user's workstation is slightly different than the rest. I was thinking of using ESX servers connected to an ISCSI San so that I can utilize vmotion. However, I can't figure out which hardware to purchase and have some questions about performance. All 90 of the vms will be used by moderately heavy users at the same time running various sql and oracle applications along with office on xp sp3. So far I am assuming 3 dual quad-core 55xx series servers with 64gb of ram each. I also like the DSN-3200-10 8x1GbE iSCSI SAN Array, but I don't know if this would work.

Would I connect 1 nic from each ESX server to a switch that is connected to all 8 of the nics on that SAN? Would this be enough or would I have to team them up 2 nics on each esx server? Would I be teaming up the 8 nics on the SAN, is a special switch required to allow this type of teaming?

How is the vmotion centrally managed? Is it installed on a seperate server or on each of the esx servers and some how linked together to one central interface?

With that being said, how many nics would I need in each of the ESX servers to allow for SAN, thin client connectivity, and vmotion?

I am open to vendor suggestions that are capable of packaging all of this together with compatible hardware.

Thanks

Health Analyzer2 can not export dataset

lincoln12 — Thu, 08 Oct 2009 14:01:27 GMT

Don't know if this is the correct forum but I did'nt find any better.

I've installed VMware Health Analyzer 2. It works fine except I can not export datasets. I get the error message "Unable to export dataset. See logfile for more info" The logfile say that it zip the files to a temp dir and then start to download the files from temp...
But there is no zipfile in mentioned temp dir. Any idéas what's wrong?

Current HA policy failing on Service Console Network

scott_k2003 — Wed, 07 Oct 2009 13:04:16 GMT

In our environment we have six ESX in one cluster on a set
of switches. Each vSwitch has a redundant connection to both physical switches
for redundancy of all networks including management / service console. In the
last four weeks we have had an issue where one of the UPS’ to which one of our
switches connect has tripped. This causes our entire cluster to become
unavailable and while some vm's remain online others do not. This looks to be
attributed to the heartbeat interval and the default policy to shutdown VM’s
when host connection is lost, since it will hit the timeout interval while the
switch reboots.

We have redundant service console connections on each
server. Vmnic0 goes to switch 1 and vmnic1 goes to switch 2. I currently have
the service console/vmotion vswitch configured in active/active mode with both
nic’s. I have been reviewing the HA documents and they talk about using an
active/passive with rolling failover policy. So my question is, if I have both
NIC’s in active/active going to two different switches, is that possibly why I
am losing service console/management access when one of the switches reboots? I
am confused because I know for a fact that if I pull one of the cables from the
san or network vswitches the vm’s automatically failover to other nic’s in the
vswitch team and service continues uninterrupted however this has not proven to
be the case with the service console connections. As such I am confused as to
why the second active connection is not keeping access running when one of the
switches fails.

Any thoughts or ideas on this would be greatly appreciated.

Thanks

Set up a VM Server and storage

Collin09 — Tue, 06 Oct 2009 22:51:32 GMT

I have been playing with ESXv4 for quite sometime now and I want to expand my VM Server setup. What would be an ideal setup for 2 VM Host server and one storage? I want to setup a disaster recovery setup between VM server. What will be the specification of host server and storage i will need?

We do not have a whole lot of budget but I'd like to present to my management a cost effective setup of VMWare.

Your input or suggestion will be greatly appreciated.

Thank you.

Collin M.

multipath script to balance I/O to a HP SAN

big vern — Fri, 02 Oct 2009 09:08:36 GMT

Howdy, all,

Running HP Blades connected via dual fabrics to active/active HP EVA 8000 SAN

Used the below script on a test esx server. It works in a fashion in that it cycles through the LUNs and changes the preferred paths from the default of the first path. It does not hiowever change the active path.

I am a complete novice at scripts - is there a way to do this , the script examples on the net seem to show that the active path changed along with the preferred path!

One other thing, is there any way to alter the script to specifically exclude 2 of the paths 9so that physcial windows boxes can use those paths in isolation ?

heres the script

for PATHS in 2 4 6 8
do STPATHS=${PATHS}
COUNTER="1"
for LUN in $(esxcfg-mpath -l | grep "Disk" | grep "has ${STPATHS} paths" | awk '{print $2}')
do
esxcfg-mpath --lun=${LUN} --path=$(esxcfg-mpath -q --lun=${LUN} |\
grep FC | awk '{print $4}' | awk '{print NR "S\t " $0}' | \
grep ${COUNTER}S | awk '{print $2}') --preferred
COUNT=`expr ${COUNTER} + 1`
COUNTER=${COUNT}
if [ ${COUNTER} -gt ${STPATHS} ]
then
COUNTER="1"
fi
done
done

on test server before running script I get.:::

--root@xxx updates--# esxcfg-mpath -l

Disk vmhba0:0:8 /dev/sdg (512000MB) has 8 paths and policy of Fixed
FC 65:0.0 5001438002b097c8<->50001fe1500b1808 vmhba0:0:8 On active preferred
FC 65:0.0 5001438002b097c8<->50001fe1500b180a vmhba0:1:8 On
FC 65:0.0 5001438002b097c8<->50001fe1500b180c vmhba0:2:8 On
FC 65:0.0 5001438002b097c8<->50001fe1500b180e vmhba0:3:8 On
FC 65:0.1 5001438002b097ca<->50001fe1500b1809 vmhba1:0:8 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180b vmhba1:1:8 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180d vmhba1:2:8 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180f vmhba1:3:8 On

Disk vmhba0:1:1 /dev/sdw (256000MB) has 8 paths and policy of Fixed
FC 65:0.0 5001438002b097c8<->50001fe1500b180a vmhba0:1:1 On active preferred
FC 65:0.0 5001438002b097c8<->50001fe1500b180c vmhba0:2:1 On
FC 65:0.0 5001438002b097c8<->50001fe1500b180e vmhba0:3:1 On
FC 65:0.0 5001438002b097c8<->50001fe1500b1808 vmhba0:0:1 On
FC 65:0.1 5001438002b097ca<->50001fe1500b1809 vmhba1:0:1 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180b vmhba1:1:1 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180d vmhba1:2:1 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180f vmhba1:3:1 On

after script I get

Disk vmhba0:0:8 /dev/sdf (512000MB) has 8 paths and policy of Fixed
FC 65:0.0 5001438002b097c8<->50001fe1500b1808 vmhba0:0:8 On active preferred
FC 65:0.0 5001438002b097c8<->50001fe1500b180a vmhba0:1:8 On
FC 65:0.0 5001438002b097c8<->50001fe1500b180c vmhba0:2:8 On
FC 65:0.0 5001438002b097c8<->50001fe1500b180e vmhba0:3:8 On
FC 65:0.1 5001438002b097ca<->50001fe1500b1809 vmhba1:0:8 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180b vmhba1:1:8 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180d vmhba1:2:8 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180f vmhba1:3:8 On

Disk vmhba0:1:1 /dev/sdv (256000MB) has 8 paths and policy of Fixed
FC 65:0.0 5001438002b097c8<->50001fe1500b180a vmhba0:1:1 On active
FC 65:0.0 5001438002b097c8<->50001fe1500b180c vmhba0:2:1 On preferred
FC 65:0.0 5001438002b097c8<->50001fe1500b180e vmhba0:3:1 On
FC 65:0.0 5001438002b097c8<->50001fe1500b1808 vmhba0:0:1 On
FC 65:0.1 5001438002b097ca<->50001fe1500b1809 vmhba1:0:1 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180b vmhba1:1:1 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180d vmhba1:2:1 On
FC 65:0.1 5001438002b097ca<->50001fe1500b180f vmhba1:3:1 On

and so on - so the preffered path is changed but not the active one.

Any help appreciated.

Capacity Planner - EM " Bad host name or server name unable resolve by DNS/NETBIOS"...

melvintang — Fri, 02 Oct 2009 07:31:32 GMT

Hi all,

I'm a newbies for using the CP tool.

After I import the server list (.csv) -> test collection . Found out few of servers was failed with EM: " Bad host name or server name unable resolve by DNS/NETBIOS". "EM: WMI failed"

(Some servers are in different IP segment; Others are in same segment)

Is that because the server did not open the port 135,137,138,139 & 445?

Anyone face this kind of issue before? Mind to share with me?

Thanks a lot!

regards,

Melvin

1 Large or Many Smaller SANs?

alan_vt — Sat, 26 Sep 2009 00:43:02 GMT

I am looking for assistance here as I attempt to gain a higher level of understanding regarding SAN architectures. Currently, I use entry level FC storage arrays with a combination of SAS or SATA drives depending upon the performance requirement. Think HP MSA 2000fc level equipment. As storage needs have grown past the limits of the unit, I've just added new units. I performed an investigation on purchasing a mid-level unit, but found the cost per GB extraordinarily high which I believe is a product of using Fibre Channel drives needed to scale to higher drive counts. If the performance of the entry level units is acceptable, what other benifits am I missing? I can see the mid-level models providing higher I/O performance, easier centralized management, more scalability, and features. With my limited knowledge this still doesn't seem to equal approximately 5 times the cost per GB. My datastores generally get setup and don't change that often. Also, I don't know how often it happens with mid-level units, but a SAN failure that causes an outage is mitigated by having spare capability on the other smaller units that allow me to restore to.

Like many of us I've needed to expand my storage knowledge upon deploying VI3 a few years ago. There seems to be a bewildering array of options, with a huge cost leap between entry-level and mid-level storage arrays. As recap time approaches this has been on my mind a lot. Is there something I'm missing?

Thanks,

Alan

Question for people using FC storage with VMware

mlubinski — Fri, 25 Sep 2009 10:21:48 GMT

Hi,

Can anyone that has his ESX hosts connected to FC storage make performance test within Linux VM to see what is his write/read speed test?

First test made from Linux Centos:

read performance: hdparm -t /dev/sda

write performance: dd if=/dev/zero of=/home/testfile bs=30M
count=512

and please update your results in here. Also if someone using Netapp FAS with NFS for his VMware could upload his the same results, I would really appreciate it.

Equalogic PS4000XV - any downsides?

J1mbo — Thu, 24 Sep 2009 10:37:32 GMT

Hi All

I have received a compelling sales incentive offer for the PS4000XV - 16x SAS 15k 300GB, 2 controllers etc etc - with a pair of Force 10 switches. I have not yet evaluted the storage system (this was planned work for later this quarter) but did have a demo in Q2 and have been steadily researching the solution since.

The system will provide storage to a 20 VM Windows server farm, supporting 100 users. All low intensity. 1 Oracle server (possibily VM, depending on license costs). Some analysis showed IO rates across the farm to be ~500 IOPS on 15 min averages. An attractive feature for us is the remote replication by adding a second unit.

Obviously being pushed by sales offers is never good, however with tight budgets this would enable us to implement the remote replication which otherwise would be outside of our budget.

The vmware project itself will commence Q1 next year, this storage would be used for physical Oracle database in the mean time.

Any and all comments on this predicament very welcome! Should we bite!?

Many thanks!

VMWare Capacity Planner - Drive Type and Media Type

capranica1 — Tue, 22 Sep 2009 21:51:24 GMT

Is there anyplace to go to find out what the Drive Type and Media Type codes refer to? When I look at my File Systems report it shows numbers for these and I haven't been able to find out what the numbers stand for.

Thanks...

EMC AX4 or HP MSA 2324 as ISCSI SAN?

rmacdonaldnai — Tue, 22 Sep 2009 14:52:10 GMT

I have recently been tasked with finding a new SAN for our 3 host Dell 2950 ESX3.5 cluster. I currently have a EMC AX150i DP loaded with 8X750GB SATA drives. This unit currently has 25 virtual machines running on it. Ive found that the SATA drives just cannot handle the io the vms are generating. Ive looked at fully populating the AX150 with 4 more 750GB drives but I honestly do not need any more space just performance. Also the cost of the drives are approx $600. Ive decided based on the cost of the drives that it would be in my best interest to put the money i was planning to use on upgrading the AX150 into a new array. Thus my dilema. Ive narrowed my search down to the EMC AX4i dual controller loaded with 12X146GB 15K 3.5" SAS drives or an HP MSA2324i dual controller loaded with 12X146GB 15K 2.5" SAS drives. Both units are coming in within a few grand of each other with warranty so price of the array does not appear to be a driving force. Upgrade costs are a factor though. Pros and cons for units so far are listed below.

AX4i-
Pros -
Familiar Mgmt Interface
EMC reliability
More widely Used
3 Year 24X7 NBD Warranty included with SAN

Cons -
Only 12 drive slots(Shelf is fully populated)
Uses slightly slower 3.5" drives
Maxes out at 60 drives
Cost to add 12 more drives later is $7260 (3k for shelf and 12X$355 for drives)
Requires additional 1U UPS for battery backed cache

MSA -
Pros -
24 drive slots in same 2U form Factor(shelf half populated)
Uses 2.5" SFF Drives which have slightly faster seek times
Maxes out at 99 drives
Cost to add 12 more drives later is $4824 ( 12X$402 for drives)
Persistant cache not vulnerable to battery backing issues

Cons -
Unfamiliar Interface
Not Real Familiar with HP Products other then procurve gear( We are a Dell Shop)
Not as Widely Used
3 Year 9X5 NBD Support extra $758. Or 3 Year 24X7X4 hour onsite $1513.00 extra( not critical as price on both units with warranty are very close)

Everything im looking at it pointing me in the direction of the HP but im just unsure on the reliability of their storage products in comparison to EMC. I may be missing something in the pros and cons so I wanted the community to weigh in on what they would recommend.

Thanks

Rich

HP eva 8000 and load balancing in VMware

bigvern2 — Mon, 21 Sep 2009 10:28:58 GMT

HP eva 8000 and load balancing in VMware

Hi, we have BL465Cs with dual post HBAs connected via 2 seperate fabrics to the EVA 8000.

We are running esx 3.5 on the blades and are experiencing disk latency so I have began looking

at load balancing.

Our storage guy says that he sees nearly all traffic arriving through FP1 on controller A on the EVA.

I'm pretty much trying to learn this on the hoof so please bear with me on these questions.

1) Does / should the EVA perform load balancing accross the FP using its own management. I'm asking because If it does would I not expect to see traffic from 2 of the FP ports (as both fabric switches have 2 connections to each controller - or are these for failover only)

2) Here's an output from one of our ESX boxes:

has 8 paths and policy of Fixed
FC 65:0.0 5001438002b0993c<->50001fe1500b1808 vmhba0:0:16 On active preferred
FC 65:0.0 5001438002b0993c<->50001fe1500b180a vmhba0:1:16 On
FC 65:0.0 5001438002b0993c<->50001fe1500b180c vmhba0:2:16 On
FC 65:0.0 5001438002b0993c<->50001fe1500b180e vmhba0:3:16 On
FC 65:0.1 5001438002b0993e<->50001fe1500b1809 vmhba1:0:16 On
FC 65:0.1 5001438002b0993e<->50001fe1500b180b vmhba1:1:16 On
FC 65:0.1 5001438002b0993e<->50001fe1500b180d vmhba1:2:16 On
FC 65:0.1 5001438002b0993e<->50001fe1500b180f vmhba1:3:16 On

I am correct in thinking that this means all traffic will go through HBA0, what does the second target number mean ? - is this the FP port number or controller (if its the fp port number which controller is it going through?) I'm confused on this as FP start numbering from 1 not 0?

cheers

Capacity Planner 2.7 won't collect CPU data for Windows / Linux hosts...?

wiewior81 — Thu, 17 Sep 2009 12:21:15 GMT

Hello guys,

I'm implementing CP 2.7 within my client's comapny. We are trying to monitor 24 hosts, both Windows and Linux (CentOS, RedHat and Fedora). The problem is that Inventory collector has found CPU data only for few hosts - the rest are marked with "X" in the CPU Inv. column.

Does anybody have any idea, why? I repeated the collection process few times, checked all permissions... other data (like RAM, disks etc.) are collected w/o any problems...

Thanks in advance,

Piotr

Midrange Storage : 3PAR, Compellent, and NetApp

meistermn — Fri, 11 Sep 2009 09:18:11 GMT

After reading on the url http://vcloudexpress.terremark.com/pricing.aspx , i look at their partners, and as storage partners http://vcloudexpress.terremark.com/ emc and 3par is listed.

Look more deep at 3par and one their performance url http://www.3par.com/products/performance.html are some points to discuss.

I was astonished that the midrange 3par has as much IOPS as the HDS Enterprise Storage Array and is 5 times cheaper than the hds. What did i miss here?

Allthough the spc-1 iops are very good. Although the responce time under 10 ms is very interesting.

The netapp solution has this metrocluster features. What does 3par and Compellent offer ?

Capacity planner fail with Redhat advanced

meriksson — Thu, 10 Sep 2009 23:25:42 GMT

hey guys

im having some issues with using Capacity planner and Redhat

I only get the amount of RAM and Disk space - everything else is missing.

The errors below show all of the 7 linux servers that are being monitored even though only 3 are having issues - seems that some rpms are missing

Has anyone seen this before?

09/10/2009 17:21:57:1:5096: Inventory.GetUnixInv: Collecting inventory for tango
09/10/2009 17:22:35:2:5096: Inventory.GetUnixInv: Importing inventory file (C:\Program Files\VMware\VMware Capacity Planner\temp\Inv_tango_2D0EB4D7-BA9D-2D4F-BCDD-A6C755DE3CA1.txt) for tango
09/10/2009 17:22:35:0:5096: Inventory.ProcessInvObjects: 119-tango Command = `lsdev` returned Error = not found in /bin:/usr/bin:/usr/sbin:/sbin:/usr/contrib/bin
09/10/2009 17:22:36:2:3124: Inventory.GetUnixInv: Importing inventory file (C:\Program Files\VMware\VMware Capacity Planner\temp\Inv_172.16.1.183_2B40B80B-39F0-3940-8AC8-96B5E11CE2A8.txt) for 172.16.1.183
09/10/2009 17:22:40:2:4904: Inventory.GetUnixInv: Importing inventory file (C:\Program Files\VMware\VMware Capacity Planner\temp\Inv_172.16.1.201_3DC637D7-5236-EA4E-ADDA-7D3BCB25A9C5.txt) for 172.16.1.201
09/10/2009 17:22:41:2:5740: Inventory.GetUnixInv: Importing inventory file (C:\Program Files\VMware\VMware Capacity Planner\temp\Inv_fccgrpwise3_95B346E8-E6BB-9F45-BC10-763A5333C732.txt) for fccgrpwise3
09/10/2009 17:22:44:2:584: Inventory.GetUnixInv: Importing inventory file (C:\Program Files\VMware\VMware Capacity Planner\temp\Inv_172.16.1.54_4C6EFFBB-7576-1645-8145-5A62FEB565D6.txt) for 172.16.1.54
09/10/2009 17:22:48:0:3124: Inventory.ProcessInvObjects: http://30-172.16.1.183 Command = `lsdev` returned Error = not found in /bin:/usr/bin:/usr/sbin:/sbin:/usr/contrib/bin
09/10/2009 17:23:07:0:4904: Inventory.ProcessInvObjects: http://33-172.16.1.201 Command = `lsdev` returned Error = not found in /bin:/usr/bin:/usr/sbin:/sbin:/usr/contrib/bin
09/10/2009 17:23:10:2:5572: Inventory.GetUnixInv: Importing inventory file (C:\Program Files\VMware\VMware Capacity Planner\temp\Inv_hydra_ECA4E859-3F6A-A746-8675-AE8A2394ACA0.txt) for hydra
09/10/2009 17:23:22:0:5740: Inventory.ProcessInvObjects: 44-fccgrpwise3 Command = `mii-tool` returned Error = not found in /bin:/usr/bin:/usr/sbin:/sbin:/usr/contrib/bin
09/10/2009 17:23:26:2:5976: Inventory.GetUnixInv: Importing inventory file (C:\Program Files\VMware\VMware Capacity Planner\temp\Inv_brahms_7B69EB46-3FF4-4249-AE1D-7A626295FFBC.txt) for brahms
09/10/2009 17:23:36:0:584: Inventory.ProcessInvObjects: http://43-172.16.1.54 Command = `mii-tool` returned Error = not found in /bin:/usr/bin:/usr/sbin:/sbin:/usr/contrib/bin
09/10/2009 17:23:50:0:5572: Inventory.ProcessInvObjects: 116-hydra Command = `lsdev` returned Error = not found in /bin:/usr/bin:/usr/sbin:/sbin:/usr/contrib/bin
09/10/2009 17:24:09:0:5976: Inventory.ProcessInvObjects: 117-brahms Command = `lsdev` returned Error = not found in /bin:/usr/bin:/usr/sbin:/sbin:/usr/contrib/bin
09/10/2009 17:24:30:1:5704: Inventory: Inventory Collection Complete for Job Task ID '14'

"Enterprise-Class" Small Business Network (10-20 users)

ConcordMotorsportNetAdmin — Fri, 04 Sep 2009 16:46:13 GMT

I posted a variation of this question on my previous discussion, but I'm interested in everyone's input ... and I wanted to clarify and update my questions after factoring in the helpful ideas and concepts I've been presented so far -- thank you all B:^) ... I've been studying the attached diagrams trying to ascertain how to scale the right VMware solution for this small 10-20 user VMware View environment, including one Terminal Server VM and reasonably-priced fault tolerance -- down to the common denominator for a small business "micro-Datacenter" ... using the minimum number of bare-metal servers ... and fitting all this into a 22" tall rack". If anyone is up to it ... I think a simple, quick, MS paint style edit of one of these network topography layouts I attached, might speak volumes to me about how this should look. I've examined the aspects of the implementation I'm still a bit fuzzy on to formulate these questions -- they pertain mostly Bare-Metal requirements:

(A) Given the small scale of this 10-20 user implementation (and very limited-yet-flexible budget), how many physical servers I should expect to require in order to decide on what to purchase for the project? ...

(B) ... Which VMware components must be installed on which of the minimum number of required physical machines? ...

(C) ... (this is the more complex question): Can any of the required physical machines (such as for dual redundant Connection Broker/Servers, or for the Windows Server for Virtual Center) ... be (separately, or running on the same PC) ... be just simple one-socket, single or dual-core desktop boxes with perhaps the added reliability of a mirrored OS disk array? -- Or would it be almost necessary or overwhelmingly preferable to get more fault-tolerant hardware for these infrastructure nodes? -- Maybe a dual-node 1U rack-mount system such as the 2-socket-per-node Transport GT28 or the 1-socket-per-node Tank GT24 ... and re-purpose the existing dual-socket Opteron TYAN Transport TA26 server (with it's 1.5 TB RAID6 array) as a NAS or SAN storage appliance + an identical physical TA26 machine perhaps, to support the High Availability feature in Sphere Essentials Plus package for the storage infrastructure?

Thanks so much for your responses,
every little bit helps!

Changing your virtual networking setup to reflect best practices

darrelleddy — Fri, 28 Aug 2009 12:53:24 GMT

I am running ESX 3.5x on two Dell PE R805 with a Dell.EMX AX4-5i array. I also have two GB switches. I want to change my virtual networking to better reflect best practices found in the VIOPS article http://viops.vmware.com/home/docs/DOC-1228;jsessionid=8EEB5BDB62598065E8ABCB55F077F36B. Since I have vm's running on the system now, can I just down the VM's make the changes to the networking and then reattach the hosts to the storage?

Darrell Eddy

Any way to enable enhanced vmotion on an existing cluster with out bringing down the hosts?

jeremypage — Wed, 26 Aug 2009 18:03:00 GMT

We have a bunch of systems running on IBM x3850m2 servers. The newest box we've gotten has slightly different processors requiring us to enable enhanced vmotion to move them. Is there a way to get these machines working with out rebooting the VMs? Maybe create a new cluster (or current one does not have resource pools defined on it but does have a lot of folders organizing the VMs and Templates.

I think we may have painted ourselves in a corner.

Vmware with Recover Point Stretch Cluster

sunvmman — Wed, 19 Aug 2009 17:07:36 GMT

EMC has a paper how to have HA using Recoverpoint Stretch cluster . However, the paper is only for Hyper-V. Does it work with vmware and if so, any paper such as this ?

http://www.emc.com/collateral/software/white-papers/h6346-disaster-recovery-geographically-dispersed-cross-site-virtual-environment-wp.pdf

NFS shares setup, and sizing for VMware

mlubinski — Wed, 19 Aug 2009 14:54:22 GMT

Hello,

I would like to ask you about suggested config of NFS volumes (shares) to make best performance for VMs placed on our VMware platform.

Our current NetApp NFS config is like this:

Cluster1:

NFS1 - 2,5TB - aggregate 1 (30 VMs)

NFS2 - 1TB - aggregate 2 (13 VMs)

Cluster2:

NFS3 - 3TB - aggregate 1 (40 VMs)

Cluster3

NFS4 - 3TB - aggregate 1 (60 VMs)

NFS5 - 1,5TB - aggregate 1 (17 VMs)

NFS6 - 1TB - aggregate 1 (9 VMDK files - data disks for VMs)

NFS7 - 1,5TB - aggregate 2 (49 VMs)

NFS8 - 1,5TB - aggregate 2 (8 VMDK files - data disks for VMs)

Cluster4

NFS9 - 2TB - aggregate 1 (50 VMs)

Cluster5

NFS10 - 200GB - aggregate1 (used for linux OS only) - 5 VMs

NFS11 - 640GB - aggregate1 (used for linux DATA only) - 5 VMDK files

NFS12 - 80GB - aggregate1 (swap) - for swap location (separate swap file vmdk)

NFS13 - 400GB - aggregate1 (Windows OS only) - 7 VMs

NFS14 - 1TB - aggregate1 (Windows DATA only) - 13 VMDK files - data disks

NFS15 - 1TB - aggregate 2 - 7 VMs

NFS16 - 1TB - aggregate 2 - 15 VMs

NetApp Aggregate1
is 78% in use (this is why we started moving VMs to second aggregate).

My other questions are as following:

1. I was thinking to split these big NFS shares to smaller one ~500-800GB (which is better for higher performance?), and
keep no more than 15 VMs on such NFS share. Should this improve overall
performance?

2. Does it matter (for performance) if I will use one NFS share for OS, and second NFS for this VM Data vmdk? Or is it ok, to keep both data/os vmdk files in one place (no difference in performance?)

3. I have doubts about config of Cluster5 with SWAP NFS share. Probably this is misconfigured, because all VMs (does not matter if windows, or linux) they have such config: Disk0 (system=8GB), Disk1 (swap=1G), Disk2 (data). I was thinking about moving all VMDK files to one datastore (OS datastore). What do you think of such config that is now?

4. Pls take a look into this drawing. Am I right, that adding one "free" NIC (marked with dot) to Storage switch SW1, SW2 etc, will not help in better network performance from ESX/VM to NetApp, right? Because I was thinking to attach this free nic to public network, to balance network load between VMs.

Thank you in advance for any answer.

vSphere and Data Protection Manager 2007

ctfoster — Wed, 19 Aug 2009 06:00:28 GMT

Hi.

Anybody out there had any experience using Microsoft Data Protection Manager 2007 with VM's. The particular scenario I'm looking at the moment is using a second DPM server in an offsite location to backup the first DPM and then using this a a source for restoring to a set of VM's at that site.

Using one DPM to backup another for DR purposes is recommended practice - but I wonder if anybody has taken it a step further and used the backups on the remote site to keep a step of VM's in step through periodic restores. Before I try messing with this it would be great to get some feedback from somebody else who has tried it.

Cheers

CTF

Life Cycle Policy For VI3

MHAV — Wed, 19 Aug 2009 05:11:01 GMT

Hey Guys I wonder what you all think about the upcoming end of the General Support for VI3. Support Statement

Do you plan on migrating to vSphere and when do start with it.

Regards

Michael Haverbeck

Starwind ISCI offering

goku — Mon, 17 Aug 2009 20:37:54 GMT

Guys I would like to know who here has experience with Starwinds ISCSI products? I'm working on a tight budget basically I need to deploy some virtual machines for a internal group they have 2 really nice Dell Poweredge 2950 servers that run 15k 400 gig sas drives. So I was thinking about using Starwinds product on those servers the servers will be running Windows 2008. So if you guys have any knowledge of the product please let me know I'm speaking in terms of reliability and performance.

Latency Benchmarks between Intel Nehalem and AMD Shanghai/ Istanbul CPU

meistermn — Sun, 16 Aug 2009 22:58:43 GMT

Came a cross a zdnet articel about micro benchmarks. The results I put in a table, to get an overview which is at the moment a better choice in cpu latency, so not only vmware vmmark results .

The resuls for the four benchmarks can be found on the following url. Clear is that the Nehalem Desktop CPU was used, not the nehelem server cpu. Although AMD'S SHanghai, not the Istanbul was used. Are there any new microbenchmarks out?

http://www.zdnet.de/zentrale_speicherung_und_rechenleistung_storage_server_in_unternehmen_virtualisierung_mit_server_cpus_leistungsbremse_inklusive_story-20000003-39199764-11.htm

	Intel Nehalem EPT	AMD Shanghai NPT
forkwait 32 bit	3,159 sec	6,525 sec
forkwait 64 bit	1,595 sec	3,879 sec
getppid 32 bit	0,480 sec	0,725 sec
getppid 64 bit	0,382 sec	0,592 sec
memsweep 32 bit	5,423 sec	13,865 sec
memsweep 64 bit	5,451 sec	15,248 sec
sum 32 bit	3,605 sec	5,520 sec
sum 64 bit	1,970 sec	3,195 sec
End Result	Intel wins all microbenchmarks

Desktop as a Service operations

yorkie — Sun, 16 Aug 2009 21:28:01 GMT

Hello all, I couldn't work out where to ask this question as it's not specifically product related, but I'm hoping a VDI customer + expert can feedback on my observations about VDI anti-patterns:

1. Liberal use of the Somebody Else’s Problem (SEP) field.
2. Indifferent business case
3. VDI changes the desktop lifecycle
4. Virtual Desktops are NOT Virtual Servers.
5. Nobody owns VDI.

More detail about what I'm on about on ViewYonder - you will find some good advice there and I would love to hear from new and old (in terms of VDI experience) alike. I'll re-post a summary of the responses, which should help everyone with their VDI operations.

Thanks in advance,

Steve

SLA Planning for virtual machines

heybuzzz — Tue, 11 Aug 2009 13:23:47 GMT

I'm working on some questions to add to our standard SLA we sent to vendors before we purchase their product. We have a section for physical servers, but we do not have one for virtualization/virtual machines. Since we mange the VM's as if they were physical servers I'm have a hard time thinking of many questions. They will never touch my ESX hosts so I'm not really sure if I even need a section for VM's.

Do any of you have SLA's for your virtual machines?

THanks

Please advise SAN / VMFS extents quastions?

MLaskowski012 — Wed, 05 Aug 2009 12:56:28 GMT

Hey guys,

Thank you in advance for your comments / advice and recommendations.

I'm not a SAN expert and don't really know much on how our SAN is configured. What makes it even harder we have our datacenter and SAN services outsourced; we are using a shared SAN not only for our different environments but also for their different clients. When it comes to my options I can ask for a RAID5 of 272GB LUN's or a RAID10 of 136GB LUN's this is all I know. We have an enterprise class EMC Symmetrix DMX 4 SAN, the way the array is configured the biggest LUN size I can get is 272GB LUN's. I really find this weird because in any mid-range class SAN I can create a 1-2TB LUN if I wanted to but I guess when it comes to the enterprise class DMX it's different. I have designed and build a huge VMware environment and it's still growing. In VMware the only way to make the Datastores bigger is using extents, I always been told to stay away from extents, and I been staying away as much as I can. I currently have about 35TB presented / 130 Datastores of 272GB LUN's plus a couple extents about 5 total datastores that are between 500-800GB. Well we are running low on space and I'm getting another 10TB presented. I know VMware support up to 256 Datastores so I'm still fine but management of storage is a nightmare at this point. We should be on ESX 4.0 by mid September, with 4.0 I know there are some options with thin provisioning, storage vmotion etc.. Storage vmotion will help me clean some of this up and get more organized. I'm getting 10TB and I'm starting to re-think the option of extents. If I get 10TB and make them 500GB-1TB VMFS volumes and storage vmotion stuff to it, delete the current 272GB LUN's and create more 500GB-1TB LUN's with extents etc... Not only this goes against everything I always been told but it's also going to the extreme from no extents to 45TB+ of extents using 500GB-1TB LUN's. I was hoping you can give me some advice.

Would you do it?

What you think about using extents in VMware?

What are the main disadvantages with doing this?

Have you seen or heard of anybody using extents to this level?

Are you aware of any gotcha's?

When building a datastore with extents does it matter what LUN's I choose to build my volume?

So what's the deal with these enterprise class / DMX SAN's they cost so much more but yet I cannot get a big LUN?

THX,

Mike Laskowski

Bladecenter Networking - optimal setup with 2 NIC ports

devzero — Mon, 27 Jul 2009 07:59:38 GMT

Hello,

i`m trying to find a solution how to hook up a IBM Bladecenter to a Core Switch environment.

I´m not a network designer, but i think i may have found a nice solution.

the problem is, that i`m unsure if this will work and some colleagues i asked have very different opinion about this.

blade1-bladeN will be ESX blades, blade 14 runs an ordinary (i.e. non ESX) operating system.

the idea is, to use VST with ESX and both nics are hooked up to the same VLANs, but by using the method described here
http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1002722
to assign appropriate active/standby adapters via portgroups. i.e. nic1 may only actively handle packets for vlan1 and nic2
handle traffic for vlan2/3 in the standard case. when some nic breaks, then one nic would handle traffic vor vlan1,2+3 alltogether.
the datasheet for the core-switches tells, that trunking over stack is supported feature.

i`m unsure if there is some logical error in this setup, as i`m not a network designer and i could have made false assumptions.

thoughts?

Do I lose vMotion on the Downgrade?

terrydavid — Thu, 23 Jul 2009 16:28:30 GMT

All.

I've been going round in circles on the site for a while now and have ended up more confused that when I stated regarding the licenses we need for a new site deployment.

The requirement is pretty simple. One site running a pair of ESX hosts, dual processors - the major requirement here is vMotion.

The second site is running a single ESX dual processor for DR purposes, no shared storage. We need to have the whole thing managed by vCenter.

The fly in the ointment is the fact that initial deployment on the host side will be ESX3.5 because of third party software. Therefore we will be downgrading the licenses for the hosts. ESXi is not an option on the DR site.

I sort of got the impression that if I purchase Advancd Edition and downgrade the license I lose vMotion - or have I got that wrong.

Any help gratefully recieved.

Raid Stripe Size For VM Backup Destination

grittyminder — Thu, 23 Jul 2009 00:43:50 GMT

Greetings!

I was wondering what the optimal raid stripe size would be for a RAID 10 NAS device being used exclusively as a backup destination for VMs? The VM files vary greatly as far as file size (e.g. 10~500KB info files, 5~500KB differential backup files, and of course, the massive 500KB~40GB image files). Right now I'm considering a 1024KB stripe size. Does anybody have any opinions/ideas on this matter?

Anyone using IBM DS8300 storage with SRM Solution?

aleph0 — Mon, 20 Jul 2009 07:42:42 GMT

Hello all,

Is there anybody out there using IBM DS8300 AND Site Recovery Manager?

Does it works?

\aleph0
____________________________
http://virtualaleph.blogspot.com/
http://vmusergroupitalia.blogspot.com/ (in italian)
###############
If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!

vSphere Upgrade Services - what are your requirements?

Daniel — Thu, 16 Jul 2009 21:05:18 GMT

What are your needs for a service from VMware or partners to upgrade from VI3 to vSphere 4? There is an existing "fixed scope" service at http://www.vmware.com/files/pdf/services/consserv-vsphere-upgrade-service-datasheet.pdf

plus VMware or partners can craft a custom service to upgrade your VI3 environment to vSphere 4.

What matters most to you? Is it

revisiting your design to take advantage of new vSphere 4 features or resource capabilities?

it's always good to rearchitect when you upgrade to a new major version (i.e. avoid stale designs?)

Have someone show and share best practices in some systems, so you can learn and deploy into your own environment as you expand?

Please feel free to add your own thoughts.

We're always looking for feedback to improve our offerings and would like to hear from you.

Thanks,

VMware Services Product Management

would you setup such environment?

mlubinski — Wed, 15 Jul 2009 18:33:27 GMT

If you have 3 esx hosts (each 2x qc cpu and 32gb ram), would you setup 6 big SQL VMs (each with 4vCPU and 16GB of mem)? If yes, how would you migrate such VMs between these 6 hosts. First problem I can see is amount of vRAM assigned - vmotion will take couple of minutes tto complete for every VM, second thing is MEM usage on hosts - in this scenario all 3 hosts will be loaded with host mem usage (each VM will then use about 12-14GB of host, showing in VC host view, right?)if so, how would you perform performance troubleshooting within OS if you think it is slower than on physical box? would you use 3rd party tools like Quest spotlight, or inside VM perfmon?

10Gb Network Performance Optimisation

f0urth — Fri, 10 Jul 2009 06:44:21 GMT

I'm running ESX 3.5.u4 on BL495 G5's With Chelsio Communications 320-BS dual 10GbE mezz Cards. ESX to ESX communication either VM to VM or ESX to ESX (via Vmotion) appears to be significantly less than I would expect.

My networking stack is a HP rebranded BNT 10gbE Blade Switch and depending on the path taken could include a Nexus 5020 backbone(if the traffic was not isolated within the blade switch, the blade switches are not directly connected(yet))

The best I have seen for vmotion is 160,000kb/sec (as reported within VC).

The best I have seen VM to VM using a socket to socket testing tool has been 70,000kb/sec. Another socket based testing tool (TCP, not UDP) Reportng 1.8gbit\sec. Still far, far beneath maximums.

Following are the ESX optimisation applied so far (Some suggested by Chelsio, other I gleaned off VMware.com articles for the 10gbe Intel cards)

esxcfg-module --set-options msi=2 cxgb3
esxcfg-advcfg -k TRUE netNetqueueEnabled
vmware-vim-cmd hostsvc/advopt/update VMkernel.Boot.netPktHeapMaxSize int 128
vmware-vim-cmd hostsvc/advopt/update VMkernel.Boot.netPktHeapMinSize int 32
vmware-vim-cmd hostsvc/advopt/update VMkernel.Boot.netMaxPCPUPktCacheSize int 256

(I'm leaving Jumbo frames out for now as VM's will not communicate via Jumbo frames, and while ESX can it is not supported under 3.5.u4)

But, to date, I'm still yet to see the value from all my 10gb investment. Any idea's?

Management / Best Practice / Config / Audit Tool

kimono — Fri, 03 Jul 2009 09:21:18 GMT

Hello Everybody,

Open ended question ... I'm keen to hear peoples opinions and recommendations on some "all encompassing" VI Management tool covering some or all of the following:

daily / weekly / monthly health checks. I am currently using powershell here with a growing number of things checked .. wondering if there's anything commercial covering the space?
best practice analyzer... once again I am using PS to validate new hosts.. Have seen tools like Tripwire Config check but how far do they go on a day to day basis? What others are out there?
Deeper issues within the environment. Looking at VCENTER for errors is one thing, but take a recent issue we had in our environment with SAN fabric, this resulted in lots of SAN path failover errors in /var/log/vmkwarning. Are there any tools that out of the box provide alerting us on this? (i know it's just central syslog, but how about smart alarming on things we care about?)
Asset Management - e.g we are currently exporting VMs custom fields from VCENTER to maintain a central repository, and then updating the central repo, and pushing back into VCENTER where needed. Then trapping new VMs that don't have this information and updating it.

Much of this falls into LifeCycle / Process space, just keen to know other large enterprises experiences with this sort of problem, so I can narrow my focus. Also I'm way out of touch with VSPHERE (training in Sept...) but have seen there's a cut down Orchestrator inside that. How have people found that - useful? Or does it scream "licensed feature only in LCM?"

Veeam Reporter seems like a fit - for some aspects - has anybody had good experiences with that?

/kimono/

Comparision Citrix Xen 5.5 and Vsphere

meistermn — Fri, 03 Jul 2009 06:28:44 GMT

Is anyone testing citrix xen 5.5 essentials versus vsphere enterprise (plus).

1.) Tests about security

2.) Which benchmarks do you use (only syntethic versus real)

Benchmark Test : Density , CPU, Mem troughput, Network throughput ,Disk throuput / IOPS, Latency test, real filecopy tests , microbenchmarks , functionality test e.g. drs ,ha, fault-toleranz, workflow creating, anymore

Capacity Planner Optimization Reports

vScott — Thu, 02 Jul 2009 17:31:40 GMT

Is there a way to filter the canned optimization reports by groups in capacity planner? I have an assessment that contaoins servers and workstations. I want to run the canned optimization reports to show only the severs or only the workstations. I can not figure out a way to filter the source machines for the reports under 'optimization reports'......

IBM Bladecenter S: internally wired SAS storage only see one vmhba

Timothy@Corismo — Thu, 02 Jul 2009 09:02:53 GMT

Hi,

We currently have a customers of ours that has implemented a Bladecenter S. It is working with HS21 xm blades that has the embedded 3i. It works beautifully but only one thing is bothering me. We set up the storage/zoning correctly. However, when we look at the vmhba's , we only see 2 vmhba's. One vmhba sees all the disks the other does not see anything. Since these hs 21 blades do not have disk inside the blade itself I would guess that one of these vmhba is the physical controller in the blade and the other vmhba is the SAS controller going to the SAS Bladechassis storage. This seems strange. Shouldn't their be 2 vmhba, one going to first bay SAS Switch and the other one going to the second SAS Switch bay like with Fibre Channel? The problem is that we do not have another customer with the same setup to compare. If anyone could confirm that is (ab)normal behaviour this would be handy.

TIA,

Timothy

How can I use the analysis from AOGCapPlan.mdb

AlanTang — Thu, 02 Jul 2009 04:36:42 GMT

Hello:

My teammate have setup the CP Data Manager and collect the data. However, they didn't register the Database & upload the data to the VMware Portal. Now, the CP machine was already housekeep. Before Housekeep, they have back the MDB file. May I know is it possible to make use of this MDB file for perform the analysis?

It was because I didn't know does this file contain the performance, inventory.... data or not.

Thanks!

I've attached a Visio of what I think my NEW environment will look like...thoughts?

theblackknight — Tue, 30 Jun 2009 00:44:59 GMT

The Visio has TWO esx hosts, but im going with 3 and all of the hosts will have dedicated HBA's.

Does it look okay, or what would you change and how?

VDR Design

steveharder — Sun, 28 Jun 2009 07:27:50 GMT

Hi,

I'm trying to separate the VDR server from our production vlan for backup procedure.

Please let me know if my design is acceptable.

I created a separate vlan (with no interface, so its completely separated from our network).

I added an additional nic to the vcenter server that is connected to this new vlan.

I moved the vdr server vnic to the new vlan by having a dedicated physical nic on the host connected to this vlan on the switch.

I have ping between the vcenter and the vdr on the new vlan.

I can't connect to the vdr via the VI on the vcenter.

I thought that the host may need an addtional SC with the new vlan id but this didn't help.

Any thoughts are welcome.

Unable to run pre-command with TSM and VCB

Fabio Pitzolu — Fri, 26 Jun 2009 07:50:36 GMT

Hello everyone.
I have a problem for a customer, which is not able to mount a vmdk for backup using the vcb framework.

The VC version is 2.5.0 build 64192.
TSM version is 5.5

The last log says the following:

--2009-06-15 14:06:57.367 'App' 12056 info-- Current working directory: C:\Program Files\VMware\VMware Consolidated Backup Framework\tsm
--2009-06-15 14:06:57.367 'BaseLibs' 12056 info-- HOSTINFO: Seeing Intel CPU, numCoresPerCPU 4 numThreadsPerCore 1.
--2009-06-15 14:06:57.367 'BaseLibs' 12056 info-- HOSTINFO: This machine has 2 physical CPUS, 8 total cores, and 8 logical CPUs.
--2009-06-15 14:06:57.695 'BaseLibs' 12056 info-- Using system libcrypto, version 90709F
--2009-06-15 14:06:57.711 'BaseLibs' 12056 warning-- SSLVerifyCertAgainstSystemStore: Subject mismatch: VMware vs ***VIRTUALCENTERMACHINE.DOMAIN***
--2009-06-15 14:06:57.711 'BaseLibs' 12056 warning-- SSLVerifyCertAgainstSystemStore: The remote host certificate has these problems:

The host name used for the connection does not match the subject name on the host certificate

--2009-06-15 14:06:57.711 'BaseLibs' 12056 warning-- SSLVerifyIsEnabled: failed to read registry value. Assuming verification is disabled. LastError = 0
--2009-06-15 14:06:57.711 'BaseLibs' 12056 warning-- SSLVerifyCertAgainstSystemStore: Certificate verification is disabled, so connection will proceed despite the error
--2009-06-15 14:06:59.180 'vcbMounter' 12056 error-- Error: Other error encountered: Not initialized
--2009-06-15 14:06:59.180 'vcbMounter' 12056 error-- An error occurred, cleaning up...
failed to prepare ***VIRTUALMACHINETOBACKUP.DOMAIN*** for backup, PrepareForBackup() returned error 1
External command failed. See error above.
Exit Code: 1

The strange thing is that, cloning the machine and launching the script, on the clone WORKS!!!
This is really strange, does someone else had the same problem?

If you need some other info let me know.

Many thanks!!

Fabio Pitzolu

Max IOPS from Equallogic PS6000e

scott_k2003 — Thu, 25 Jun 2009 11:32:15 GMT

What are the max iops one can expect to get from the from the Equallogic PS6000e?

Planning backups for Lotus Domino virtual machinea

caddo — Thu, 25 Jun 2009 08:08:41 GMT

In this scenario the customer is using a combination of Backup Exec, VCB and vRangerPro to backup the VI.

My question is: is it possible to make consistent backups of Lotus Domino databases using the above tools considering that the customer has no Backup Exec agents for Lotus Domino? Will the snapshots taken by VCB/Backup Exec or by vRangerPro maintain consistency of Lotus Domino DBs?

The simplest way to do this would be to use VCB integration of Backup Exec and save all .nfs file to tape along with all the rest of the data that is already being saved this way (mainly file server data); now, since Domino is not VSS enabled i am concerned to suggest a way to give them consistency in the backups, i don't want to find out later on that you can't restore the DBs because they are corrupted or not up to date.

vRangerPro is used mainly as a disaster recovery tool backing up full virtual machines without VCB integration.

Hope someone has experience in this scenario. Thank you.

Are you planning for a large VM deployment? We are and would like to talk with others doing the same.

Marc Gelinas — Wed, 24 Jun 2009 20:43:12 GMT

Hello all,

We are in the process of looking at how to plan for a redesign of our existing Lab Manager environment to a staged scalable deployment of Lab Manager/vSphere4.

Here is where we are at today,

Equipment:

Dell 2950 x 5

Dell MD3000i (15 x 300GB SAS) (RAID 10)

Dell MD1000 (attached to the MD3000i) (15 x 450GB SAS) (RAID 5)

What we are considering for upgrade options:

OPTION 1 (0-12 months usage) Support for up to 100VMs?

Adds & Upgrades:

UPGRADE - Completely reconfigure all MD3000 & MD1000 storage arrays and datastores within VC

ADD - Dell R710 x2 (Dual proc Quad Core, 64GB RAM, 6GB NIC ports)

UPGRADE - Dell MD3000i (UPGRADE DRIVES to 15 x 450GB SAS) (RAID 10)

UPGRADE - Dell MD1000 (attached to the MD3000i) (15 x 450GB SAS) (UPGRADE to RAID 10)

UPGRADE - Replace 2 existing Dell 5324 network switches with Cisco 3750's

OPTION 2 (12-24 months usage) Support for up to 200 VMs?

Adds & Upgrades:

REMOVE - MD3000i and MD1000

ADD - 2x Dell EqualLogic PS6000E storage arrays

ADD - Dell R710 x2 (Dual proc Quad Core, 64GB RAM, 6GB NIC ports)

ADD - 2x Cisco 3750 Network switches

OPTION 3 (24-36+ months usage) Support for up to 400 VMs?

Adds & Upgrades:

ADD - 2x Dell EqualLogic PS6000E storage arrays

ADD - Dell R710 x5 (Dual proc Quad Core, 64GB RAM, 6GB NIC ports)

We would love to know what others in similar shoes are doing. We don't have anything to compare against. Please post any feedback you might have on any of the information above.

We are particularly interested in what hardware others have deployed in support of similar amount of virtual machines.

Some help with vCenter Server RAM usage

evolmar — Wed, 24 Jun 2009 16:50:06 GMT

Hi all,

I'm having some on going issues with my vCenter server. Nothing serious I don't think but something I'd like to get hammered out.

It's currently residing on a Windows Server 2008 64-bit VM. I have 4.5GB assigned to it. The problem is that it is always consuming most of the RAM allocated to it. It's always triggering yellow or red alarms. Originally it was set to 4GB. I added another 500 MB hoping that will solve the issue to no avail. I'm convinced I'm doing something wrong with this implementation. Attached is the resource allocation tab on the vCenter server.

This is not a large environment. Only one host 5 VMs. If anyone can give me a couple of pointers that would be much appreciated.

Regards,

e

Capacity planner performance summary reports

DHaley — Fri, 19 Jun 2009 19:49:10 GMT

I used to be able to pull up performance summary reports that told me the industry average as well as averages of the environment I was monitoring. Now it does not show that, it only shows something when I select group by systems. See attached

HP Bladesystem Virtual Connect/ Extreme Networks Swtich/ 802.3ad with Vmware 3.5 networking questions - please help!

digitalex — Wed, 10 Jun 2009 13:51:33 GMT

We are currently having a issue sometimes on our switch where the grouping of the switch ports go into a Not Distributing state on the Extreme networks switch. After doing some research I am finding that Vmware or the HP Bladesystem Virtual Connects do not support 802.3ad LACP with dynamic but with static only. Is this true? Is there anyone out there that knows Extreme Network switches and how to configure the trunking for Virtual Connect on HP and the VMware ESX servers before I have to call Vmware support?

Extreme switch config:

enable sharing 1 grouping 1-8 dynamic

Vmware ESX config:

Active Adapters 1000 Full "Route based on the originateing virtual port ID" load balancing, Link Status only

My current setup for info:
Virtual Center 2.1

ESX 3.5 up1 on 3x HP DL460c G1 16gb
EVA4100_1
20 Windows 2003 Ent VM's

Vi3 Enterprise + vSphere 4 Offsite ... need help with completing the solution in BP

ian.wills — Tue, 09 Jun 2009 21:29:11 GMT

We've been migrating from physical to virtual servers over the last 14 months, and are nearing completion of our final p2v. Now, we need to focus on a solution to account for offsite failover, offsite online snapshots of VM's and datastores, and offline backups. What I'm looking for is some best practices for snapshot management, offline backup, and general performance tips...

We run an online school, which necessitates a fairly high requirement for uptime. We have secured a location and some hardware to establish a DR site, but one of the things I'm quickly becoming overwhelmed with is the configuration and design for the DR strategy. I'll outline our current hardware / software configuration below:

Site A:

VCenter Server Enterprise (VM) - HA, DRS enabled w/ EVC

Equalogic ASM / VE Console (seperate Win2k8 VM)

2x Dell PE1950 16GB SAS Boot w/ QLE4062 and Pro1000MT adapters ESX 3.5.0 b110268

2x Dell R300 4GB SATA Boot - onboard NICs ESXi 4.0.0

Dell Equalogic PS5000E dual controller - 8x 1TB Half Populated - New VMFS Datastore + RDM for server data volumes

2x D-Link DXS-3227 L2+ Switches w/ STP disabled, broadcast storm control disabled, jumbo frames enabled, flow control enabled

Synetic OEM 16-Bay ISCSI SAN - Original VMFS Datastore

100Mb/s Full Duplex Fiber to Site B over OpenVPN.

Virtual Machines:

Win2k3 -Firstclass Server (300GB)

Win2k3 - Net/DHCP/DNS/AD

Debian 501 - Web Apache + RDM

Debian 501 - File Server (netatalk)

Debian 4 - OpenLDAP server

+ more etc

Site B:

VCenter Server 4.0 Standard (ESXi Management Kit) (Win2k8 VM)

Equalogic ASM / VE Console (Separate VM)

2x Dell R300 4GB USB Boot ESXi 4.0 - onboard NICs

Dell EQL PS5000E dual controller - 8x 1TB half pop - DRSite VMFS Datastore

D-Link DXS-3227 switch

Site A has semi-frequent power outages that makes it less than ideal for an internet hosting facility, but the powers that be insist on our primary servers being stored onsite, with an emergency failover DR site in the case of an environmental disaster. To fight power outages, we installed enough UPS backup to keep the Vi3 cluster up for about 600 minutes. (A generator was viewed as too costly...)

Currently, the two PS5000's are replicating the VMFS volume from Site A to B, but it is not remounted on Site B. We looked at SRM, but again, costs became an issue... Ironic, I know.

My primary goal is to establish an automated means to replicate VM data to the DR site, and in addition backup VM's and server data to an offline storage (the older Synetic iSCSI SAN, will become the offline data store for archives, in lieu of a tape solution)

I'm sure this initial coverage is poor, but any thoughts or questions to help flesh this out?

Thanks so much...

esxi and dell 710 Server

trinimoses — Tue, 09 Jun 2009 16:59:06 GMT

My vm budget was cut down DRASTICALLY and i was only able to get a dual socket, quad core dell 710 with 8GB ram and 2 160gb hard drives....

I am looking at installing the free download of ESXI version 64bit and run on that dell server 2 vms of windows 2008... on it.

MY question here is .. how good is esxi compare to the enterprise product ?

Does anyone have any video tutorials of installing and controlling a windows 2008 server on esxi ??

Disk alignment questions with Esx 4

evolmar — Mon, 08 Jun 2009 22:30:04 GMT

Hi all,

I want to begin by saying I apologize in advance if this has been posted numerous times. I've tried numerous searches in here and google and the various white papers to no avail.

We have a Clariion CX3-40 and have mapped a 525GB LUN to our very first production ESX host. The host has ESX 4 installed.

Within the host I see the LUN and have formatted it with VMFS with 8 meg block size. All is well until I I realized I have to address alignment concerns. When one used the VI client to create the datastore,does vSphere automatically handle the disk alignment for you? What about guest OSes?

Also regarding the 8 meg block size. I read here that using 8 meg block size right off the bat is the way to go. True?

Thanks in advance for all of your guidance.

Regards,

e

New Install - Version 3 or 4?

tracylbrown — Fri, 05 Jun 2009 03:46:44 GMT

We have recently purchased some Dell 710 servers and EqualLogic PS6000 arrays for our VMware environment. We have been running some free ESXi 3.5 servers for several months and have been very happy. Now that we have our hardware and software purchased. Should we just go ahead with a version 3 setup or go ahead and make the dive into version 4? Being we have no background to work from on past VMware upgrades and version releases, I would like to get some feedback from some veterans on this one.

Thanks ahead of time for any feedback. It will be appreciated.

Thanks.

--tracy

Searching for indepentend "rapid cloning" tool for templates

meistermn — Mon, 01 Jun 2009 15:01:33 GMT

Come across this nice video about netapp's rapid cloning tool.

http://www.youtube.com/watch?v=E7lccBlE5cw&feature=channel_page

I like this tool, but what about an storage and server vendor independent "rapid cloning tool".

If you have 3 different storage vendors in house, than it makes sense to have an independent "rapid cloning tool".

I would like to see in this tool management of ip address and computernames.

Although the deployment of the template should not use pxe and dhcp ( dont like px/dhcp errors and configuring relay dhcp agents -> simplefy this process).

An onther point to replicate/rollout the golden template from the central location to the branch offices.

The dashboard on the server says is 42 systems found, 41 systems consolidatable

Jeff07 — Sun, 31 May 2009 10:17:33 GMT

I have VMware capacity planner running for a client.

The dashboard on the server says is 42 systems found, 41 systems consolidatable, Collector registered, data sent.

When i go to the Vmwar'se website and look at the reports its says "1 system ready for review".

The system count uploaded is 47 systems but i get no details in the graphs or reports about all the other systems uploaded.

Any ideas why? added screen shot

specifying date range for CP assesment

odiaz_iisl — Thu, 21 May 2009 19:02:14 GMT

My customer had a cp asessmen done last year and there were 12 weeks of data collections during two differnt time periods. They are only interested in doing an assesment based on the perfromance metrics for one week within each time period. Is there a way perfroma consolidation assesment based on a given date range?

Upgrade path from acceleration kits?

LiamCurtis — Thu, 21 May 2009 16:28:30 GMT

Hello all. Faced with choosing a virtualization platform and my head is spinning trying to compare Hyper-V, Xenserver, and VMWare. Bottom line is I would go with VMWare in a heartbeat if cost were not an issue, so I am looking at the acceleration kits.

The problem is...what happens when you outgrow the vcenter 'f3 server limit? Is there any sort of upgrade path?

Sizing - Growth of Existing ESX Datacenter

CraigAlexander — Tue, 19 May 2009 08:09:28 GMT

Hi,

We have 12 ESX 3.5 Hosts in our Datacenter across 3 clusters of 6, 3 and 3 nodes respectively.

The Hosts are all Dell 2950 servers with 2 by Quad Core CPUs and 32GB of RAM.

These hosts are currently averaging about 20% CPU and about 70% memory used.

The hosts appear to be performing OK but I am concerned about the effect if a Host in a cluster has a failure and all of the VMs are started up by HA on the other nodes in the cluster. I am concerned that the memory use on the other hosts will increased beyond the physical RAM of those hosts and cause performance problems to some or all of the VMs running on those hosts. I would also like to able to easily vacate all VMs from a Host in the case of maintenance on the server such as upgrades without effecting other hosts.

Does anyone know of any guidelines or 'rules of thumb' on how much RAM should be allocated as total of the VMs on a host to avoid the concerns above.

Or should I just set the following in the cluster setting and let VMware decide when I have enough VMs

Number of host failures the cluster can tolerate 1

Prevent VMs from being powered on if they violate availability constraints checked

Regards

Craig

Are you moving to 10 gigabit ethernet?

heybuzzz — Thu, 14 May 2009 12:54:29 GMT

I'm curious to know if anyone out there is running (or plan on moving that direction) 10 gigabit Ethernet to their ESX hosts. I had a talk with one of our local Cisco reps yesterday and he was telling me that I could replace my six pNICs (2 - SC/VMotion and 4 - Production) and dual 4 Gbps FCcard with just two copper cables running 10 Gig E. Going from 6 different physical cables from 3 different NICs to two different switches in my data center gives me a good piece of mind. To take that all away and just rely on two cables...hmmmm. I was called into this meeting when they started to talk about how great it was and would be a cost saver. Anyways, I have done some reading on my own and would like to hear some personal experiences. Thanks

Pretty good read... http://virtualgeek.typepad.com/virtual_geek/2008/06/10-gigabit-ethe.html

How to delete the collector information for a CompanyID in the dashboard

puzzledtux — Thu, 14 May 2009 06:34:19 GMT

Recently completed the capacity planning training and am testing some things in the CP dashboard. I would like to know if there is an option in the dashboard to delete the collected information during a CA (detailed reporting) for a CompanyID.

I would like to know this because while testing can I send request to vmware to create new test company ids to have a complete assesment done? or else can I delete the existing collected data for my company id and start again from there only.

ESX with netapp NFS metro cluster

stm_cro — Wed, 13 May 2009 11:52:59 GMT

Hello,

I had always wished ESX would manage LUN mirroring just like any other operating system does (The reason is that when you host an application that should NOT stop, you want an easy and automatic failover in case of computer room failure, or storage array failure).

VMware DRP solution looks really heavy to me compared to a windows/linux/HP-UX metro cluster (with a storage volume manager like Veritas).

Then, I realize that NetApp provides a NFS metro cluster that look like a perfect storage solution for higly resilient VMware infrastruture.

My problem is that I don't see anyone talking about this solution. Am I missing something (beside the price)?

Is it as nice as it looks? Has anyone used/setup NFS metro cluster with VMware? Any usefull link/whitepaper?

Thanks for your help,

resizing vmdk? (smaller)

wb2 — Tue, 05 May 2009 18:30:00 GMT

Is there any easy way to resize a Windows VM's disk size as well as the associated vmdk? I made my template to be 100gb, but I want to be able to roll out 50gb disk sizes as well.

Error Collecting Perfmon Data with CP

JimYanik — Fri, 01 May 2009 22:22:26 GMT

On several W2K3 servers I can't get perfmon data. DNS and NetBIOS name resolution work fine but when I try to collect perfmon data I get "Failed"

When I test connection, I get the following;

Services Perfmon Failed Unknown Error(the specified object is not found on the system)

Any ideas?

Capacity Planner CA Reports

rcheeseman — Fri, 01 May 2009 18:39:20 GMT

I have a question regarding the CA reports. I am actively working on 4 different customers Capacity planning. I understand the interim and final reports for the Capacity Estimates, but my problem is with a CA reports. I have went through all of the dashboard, and ran all kinds of reports and scenario's and it allows me to view report, but I was looking for it to compile the 30+ page report for the customer (Like the CE report, but larger and with more info. Do I have to export each report and scenario into Excel and build my own? I thought the CA report could be downloaded as a pdf like the CE report. What am I doing wrong.

Thanks

Ryan

Set default power-on boot delay

sbartle — Fri, 01 May 2009 02:33:56 GMT

Hi all,

I would like to know if there is a way I can change the
default Boot Options - “Power-on boot
delay” = Delay 0 ms to a different value site wide? (Delay = 5000ms) I know this can be done by
a script for current guests, but I would like to make this the default setting for
all new guests. Is there a template that the default build uses for these
initial settings?

Thanks

Backup Options

UoGsys — Wed, 22 Apr 2009 21:35:59 GMT

I'm keen to hear other members opinions of the following products that plug-in to VCB. We're looking to replace our backup system in the next few months so we have a blank canvas when it comes to protecting virtual servers.

vRanger - I wasn't massively impressed here as the current version seems to have fallen behind it's competitors in terms of features. These seem to including items such as file level restore (mounts the entire vmdk file), lack of data de-duplication to name a few. I was assured when I visited Cannes though that these would be available when the new version is released.

Veeam Backup - Seems to offer all the above features as well as de-deplication and proper file level restore. Plus 50% off at the moment.

VCB Plug-in for BackupExec 12.5 - I know very little about this backup proxy, being pushed to us through our backup consultation project and will obviously integrate with our full backup solution. Does anyone have any positive/negative comments regarding the use of this product. I'm guessing it will offer a backup to tape option, not present in the other products?

VM Data Recovery - The new boy on the market, we'll be moving to the enterprise version of vSphere so will involve no cost. I was impressed with the overview and demo at Cannes, but the current drawback is that file-level recovery is an experimental feature.

I'd love the hear your thoughts on the above products and any significant features I've missed to help us to make a decision!

Application Compatibility

rbmadison — Wed, 22 Apr 2009 17:56:17 GMT

I know VMware's big on getting most if not all apps to run on VMware. Are there any VMware resources or information on what applications have issues and if VMware is aware and is working on improving compatibilty with specific applications? We are trying to run an application that uses intersystems cache and they have a document stating cache doesn't work well with VMware. Anyone that may have suggestions or experience with cache on VMware or experience in working on application compatibiliy with VMware that would be great.

Rick

Virtual Desktop deployment and DHCP

prutter — Mon, 20 Apr 2009 15:25:25 GMT

I am planning on deploying VM's for desktops and wanted to test before i push it out. The question i have relates to DHCP and 2 subnets. As it stands now my current subnet 1.1 is 80% full so in order for me to deploy 30 more VM's i will need another subnet i assume. Can this be done with a virtual switch or does the routing need to be done with a physical switch. Any help is appreciated.

ESX 3.5 U3 and Chelsio N310E -SR+ cards

kpeuramaki — Sun, 19 Apr 2009 00:29:09 GMT

I'm trying to install some N310E cards on an HP DL380G5 running ESX 3.5. I have the drivers from Chelsio installed on the system but they just don't load. Any ideas? p.s. - I'm tryng to do this while the physical system is about 8 hours away and I am connecting through iLO.

Thanks for any ideas!

VMware, Netapp 3140, iSCSI, NFS, Extreme Networks SAN Setup

DobsonConsulting — Fri, 17 Apr 2009 15:41:29 GMT

I just purchased a Netapp 3140 with dual controllers and 8 Gig-E ports. The plan is to use NFS for VMs. Each filer head will have a Gig-E connection to each SAN or iSCSI switch (Extreme Networks). We have a total of 10 Gig-E ports on our ESX boxes to carve up. I am used to EMC and Fibre Channel so I had a few questions on the setup.

I plan the ESX network config to be something like:

vswitch0 - SC (vlan1), VMotion (vlan2) - 2 physical NIC connections to redundant core LAN

vswitch1 - VMProd (vlan3) - 4 physical NIC connections to core LAN

vswitch2 - NFS (vlan4) - 2 physical NIC connections, one to each iSCSI SAN switch

The other 2 ports are going to be reserved for iSCSI, say vswitch3.

Now, for vswitch2 (NFS) to have the best bandwidth and failover, should I create 2 IP connections to the vswitch? How would I balance load across the different iSCSI switches in this configuration? I know there are multiple ways of doing it. If I have some NFS targets presented on one SAN and some on the other, would that load balance the traffic?

I have read many posts on this, but cant seem to find a good answer. For NFS using Netapp, what is the best way to load balance traffic? What is the best way to configure it on the ESX and Netapp side? Is it best to have 2 physically and logically seperate iSCSI SANs for redundancy or would it make sense to connect them? Is one IP or SC good for my NFS vswitch? Should I add more pnics to it?

I appreciate it.

Experience with HP BL490c: Nehalem-based blade server

WillGreen — Fri, 17 Apr 2009 10:53:00 GMT

Has anyone run any HP BL490s with ESXi yet? They're similar to the Opeteron-based BL495s, but the Intel Nehalem processors seem very promising performance wise.

Details here: http://h18000.www1.hp.com/products/servers/proliant-bl/c-class/490c-g6/index.html

I am currently trying to obtain an evaluation 490, but if anyone has already got their hands on one I'd be grateful for any thoughts or issues encountered.

Lefthand, ESX and VCB using MPIO

terrible_towel — Thu, 16 Apr 2009 18:57:49 GMT

Hi Everyone.

We are using LeftHand for our backend storage. The ESX servers use
iSCSI for access to the volumes on LeftHand. These volumes are
formatted with VMFS from the ESX servers.

I have been using a windows box with the Microsoft iSCSI initiator in
conjunction with VCB to backup the VMs (SAN to windows box directly
using iSCSI).

I have had some intermittent connection issues with the iSCSI initator to the LeftHand.

I found a paper on LeftHand support about not using the Microsoft DSM
for MPIO with the initator. I had not been using MPIO before this.

So I uninstalled the initiator and then reinstalled it, making sure not to check the Microsoft MPIO support box.

I then installed the LeftHand DSM for MPIO from lefthand.

Well to make a long story short- within an hour, 70+ VM's had corrupt
vmdk files. I had to resort to creating new volumes on the lefthand
and restoring all the VMs from their backups.

My question for anyone familiar with Lefthand, MPIO, and ESX is: is
the lefthand DSM MPIO driver compatible with VMFS file systems ? Or
does it expect only NTFS file systems ? That may at least explain what
happened.

Thanks,

Paul

Capacity Planner - I/O (Trans/Sec) stumper!

hpuxman — Thu, 16 Apr 2009 05:19:26 GMT

Has anyone seen or know why capacity planner reduces the total disk transfer I/O (of certain systems) from the existing source systems to the new recommended scenario? I don't see how virtualizing reduced the need on the disk i/o generated from the application

For example, the total of all systems in a scenario is shown below with 1410.91 being the Disk I/O transfers/sec

Capacity

Utilization

Processors

Memory

Disk

Network

Physical

Processor

Memory

Disk

Network

System Name

Make/Model

Count

Speed
(MHz)

Size
(MB)

Size
(GB)

Count

Speed
(MB/sec)

Rack
Units

Weight
(lbs)

Power
(W)

Thermal
(BTU/hr)

% Used

Queue
per CPU

% Used

File Sys
Cache (MB)

Page
File %

Paging
(Pg/sec)

I/O
(Trans/sec)

I/O
(MB/sec)

Speed
(MB/sec)

All Systems

106.8 GHz

73.5 GB

6.50

42.1 GB

876.66

9.2 KW

2.8 Tons

4.23

0.18

55.02

3,993.54

2.41

457.70

1,410.91

87.57

1.67

Then in the generated, recommended scenario it goes to 944.13.

All Systems

51.3 GHz

98.3 GB

3.504 TB

12.0 GB

123

1.2 KW

0.4 Tons

6.55

0.24

40.74

3,901.77

1.26

301.30

944.13

66.16

1.64

Multiple hosts to a single Buffalo iSCSI Server

U235 — Tue, 14 Apr 2009 21:18:07 GMT

I wanted to run this by you before I get all the hardware and realize it doesn't work.

I have two ESX Servers that I want to connect to a single iSCSI device. I don't need much performance as they're a combination of archive and test virtual machines. Anyways, I want to have access to the same iSCSI LUN from two ESX hosts.

I know VMFS supports clustering and can handle this, but I'm not sure if the Buffalo is ok with VMFS. I'm assuming it is, but you never know since it's such a low-end device. Buffalo supports formatting multiple partitions and allowing access to each partition, but I'd prefer to not have to manage two separate permissions and just have one large iSCSI data partition.

I'm not worried about data corruption since the 2 ESX servers will never be running the same virtual machine and I think VMFS will file-lock it anyways so there's a built-in precaution anyways. Buffalo says they don't support multiple host access, but they're talking more from an NTFS point of view and I agree with that, it's too dangerous to risk, but with Vmware it's a different situation.

What do you guys think?

Architecture direction for 2009 : Intel Nehalem EP 2 Socket versus AMD Shanghai 4 Socket versus AMD Istanbul 4 Socket

meistermn — Mon, 13 Apr 2009 13:46:16 GMT

I started this discussion to get out technical differences for Intel's Nehalem and AMD architecture, which include AMD Istanbul (comes 2009), but exclude Intels Nehalem EX, comes 2010.

Today we only use 4 socket Server. Most are AMD Barcelona , but although Intel Tigertons. So all Quad Core.

What are the differences for one over the other architecture:

2 Socket , - 4 Socket , number of dimms, downgrade of memory speed , scheduler, cpu queueing, hyperthreading (are thread 1 to 8 equal to thread 9 to 16), vmmark results.

CPU Part	Intel Nehalem EP System	AMD Shanghai	AMD Istanbul
VMmark result	23,46	20,50	Expected 30
Sockets	2	4	4
Cores per Socket	2 x 4	4 x 4	4 x 6
Threads	8	16	24
Hypertreading	8	-	-
Threads Total	16	16	24

Caches

	Intel Nehalem EP System	AMD Shanghai	AMD Istanbul
L1 cache size (max)	32 KB D-cache + 32 KB L1 instruction cache per core	64KB (Data) + 64KB (Instruction) per core	64KB (Data) + 64KB (Instruction) per core
L2 cache size (max)	256 KB low latency L2 cache per core	512KB per core	512KB per core
L3 cache size (max)	8 MB shared L3 cache	6MB (shared)	6 MB or more
SIMD Instruction Set Support		SSE, SSE2, SSE3, SSE4A	SSE, SSE2, SSE3, SSE4A

AMD Infos http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_8796_15225,00.html

Intel Nehalem ?

Memory Controller	Intel QPI		AMD Hypertransport	AMD Hypertransport
Version	1		HT1.1	HT3
Memory frequencies	800/1066/1333		677/800	800/1066/1333
DDR Technology	DDR3		DDR2	DDR3 with Fiorano platform
QuickPath Connect bandwidth per CPU	25,6 GB/s	Maximum Memory Bandwidth 4P System	51.2GB/s	higher
?	?	Maximum I/O bandwidth with 4P System	32.0GB/s	higher
Maximum QuickPath Connect bandwidth	51,2 GB/s	Maximum Total Bandwidth 4P System	83.2GB/s	higher

Not finished yet.

Benchmark a ds4300 for Virtual Infrastructure use

caddo — Wed, 08 Apr 2009 16:29:42 GMT

Hi everyone,

i would like to know how can i benchmark my current disk array; i would like to stress it up a bit and get some metric so that i can decide if i can use it ad a VMFS datastore and for what kind of VMs.

Can u suggest any tool/software that i can use for that? Free if possible :)?

Thanks in advance

queuestion about HA

jerry2096 — Fri, 03 Apr 2009 15:10:31 GMT

I have a question about HA without using DRS: How to decide which ESX host to failover for the VM guest on the isolated machines?

For example: we have 3 ESX hosts - ESX01-ESX03.
If ESX01 has 10 VM guests running, and ESX02 and ESX03 only have 5 guests.
1. When ESX01 is being isolated. How do the 10 VM being powered on on other 2 ESX hosts? What 's the Logic behind it for selecting the ESX host to failover?
2. If both ESX02 or ESX03 have capacity (resource of memory and CPU, mainly Memory) to host just another 5. Will those 10 VM on ESX01 being separatedly powered on those 2 ESX hosts? (We are not talking about DRS, DRS is not enabled.)

thanks

Start up issues after complete shutdown

pdrace — Fri, 03 Apr 2009 12:25:16 GMT

Due to migration to a new computer room we have had many planned and one unplanned complete power down and power up.

Our virtual center server and database server that VI uses are vms. We've had issues getting vms up and running again after powering the hosts back up. When I attempt to attach to a host using the VIC it times out. The first time this happened was after we did a storage migration. I called support and we thought we figured out that the issue was that the domain controllers used for authentication weren't available as they are virtual. I disabled the AD authentication on the host and I was able to connect to the host and life was good.

Two times after this I haven't been able to attach to an ESX host with the VIC even after disabling AD authentication. We just keep getting "server is taking too long to respond". I've had to use webaccess to start the VC and database vms for the last two shutdowns. All the hosts are in a cluster but HA and DRS were disabled before shutting the hosts down.

Has anyone else experienced this and come up with a procedure/solution to avoid this pain?