VMware Communities : Unanswered Threads - Performance & VMmark

Enable trusted execution

aRandomPerson — Sat, 21 Nov 2009 13:09:43 GMT

I have a Dell Optiplex 755, I updated my BIOS settings to Enabled Trusted Execution, VT enabled and virtualization enabled. Now my machine will boot to a blinking cursor. I am unable to get back to the BIOS? Is there a way to resolve this?

error durring javaserver workload result collection

simonpha — Sat, 21 Nov 2009 04:16:28 GMT

Hi I've run into this problem with the javaserver workload when completing the workload run. I receive the following error in STAX. It seems to me that the workload completes correctly but when trying to do the result collection it errors out.

20091119-18:55:02 Start JobID: 2, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local, Function: Main, Args: { 'CONFIGFILE' : r'C:\vmmark\VMMARK.CONFIG' }, JobName: smalltest1
20091119-18:55:02 Info Holding block: main
20091119-18:55:02 Info Received RELEASE BLOCK main request
20091119-18:55:02 Info Releasing block: main
20091119-18:55:13 Start Testcase: VMmark Setup
20091119-18:55:36 Stop Testcase: VMmark Setup, ElapsedTime: 00:00:23
20091119-18:55:36 Start Testcase: VMmark Wkld Setup
20091119-18:57:54 Stop Testcase: VMmark Wkld Setup, ElapsedTime: 00:02:18
20091119-18:57:54 Start Testcase: VMmark Run Tiles
20091119-19:53:24 Stop Testcase: VMmark Run Tiles, ElapsedTime: 00:55:30
20091119-19:53:24 Start Testcase: VMmark Results Collection
20091119-19:53:25 Error STAXPythonEvaluationError signal raised. Terminating job.

===== XML Information =====

File: c:\vmmark\xml\javaserver_functions.xml, Machine: local
Line 322: Error in element type "call".

===== Python Error Information =====

com.ibm.staf.service.stax.STAXPythonEvaluationException:

Python object evaluation failed for:
'Info: COPY FILE C:\javaserver_%u.stdout TOFILE %s%sjavaserver_%u.stdout' % (tileserver, gResultsDir[0], gFsep, tileserver)

Traceback (innermost last):
File "<pyEval string>", line 1, in ?
ValueError: invalid literal for __int__: javaserver0
===== Call Stack for STAX Thread 1 =====

[
function: Main (Line: 106, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local)
sequence: 16/22 (Line: 128, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local)
testcase: VMmark Results Collection (Line: 507, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local)
sequence: 1/2 (Line: 508, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local)
iterate: 1/5 0 tilelist (Line: 510, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local)
sequence: 1/1 (Line: 511, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local)
iterate: 2/6 JavaServer gWORKLOADLIST (Line: 513, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local)
sequence: 2/2 (Line: 514, File: C:\vmmark\xml\vmmark_main.xml, Machine: local://local)
function: GetResultsJavaServer (Line: 227, File: c:\vmmark\xml\javaserver_functions.xml, Machine: local)
sequence: 8/12 (Line: 234, File: c:\vmmark\xml\javaserver_functions.xml, Machine: local)
]
20091119-19:53:25 Info Terminating block: main
20091119-19:53:25 Stop Testcase: VMmark Results Collection, ElapsedTime: 00:00:01
20091119-19:53:25 Status Testcase: VMmark Run Tiles, Pass: 1, Fail: 0, ElapsedTime: 00:55:30, NumStarts: 1
20091119-19:53:25 Status Testcase: VMmark Setup, Pass: 1, Fail: 0, ElapsedTime: 00:00:23, NumStarts: 1
20091119-19:53:25 Status Testcase: VMmark Wkld Setup, Pass: 1, Fail: 0, ElapsedTime: 00:02:18, NumStarts: 1
20091119-19:53:25 Status Testcase Totals: Tests: 3, Pass: 3, Fail: 0
20091119-19:53:25 Status Job Result: None
20091119-19:53:26 Stop JobID: 2

Anyone have any idea what this may be?

Thanks in advance

Virtual Desktop Performance issues - ESX 3.5 VM View

ciberweb — Wed, 04 Nov 2009 17:26:43 GMT

Sorry, wasn't sure what category to submit this under.

I have a small environment running 1 VM server and 5 VM Desktops. Performance on the Desktop side is the issue.

Server Hardware:
1 - Dell 2950 / 2-Quad 2.8ghz Proc / 16gb RAM / 1.2TB of local storage - 15k RPM HD's / 2 - Dual NIC's (1 NIC = 2 VM Desktops, 1 NIC = 3 VM Desktops and 2 NIC = Server)

Software:
Server -> SBS 2008
PC's -> Windows XP Pro SP3 - Office 2003, ERP Application, File Maker, Adobe and WinZip

Virtual Hardware:
1 2.8ghz Proc / 1gb RAM / 25gb C:\ drive (15gb free)

Physical Hardware used for Virtual PC's:
Dell Dimension 2400 -> 2.4ghz Single Proc, 512mb or 768mb RAM, 100mb NIC

Issues:
VM users are seeing slowness issues when attaching documents to emails in Outlook.
Refresh is slow when changing from one email to another.
Working in one application and then toggling to another hangs the desktop for a bit.
Generally speaking applications seem to be slow.

What is my bottleneck? Is this typical or is there a config change that can be made to resolve the issues?

Help would be greatly appreciated.

Marc

creating vm pc's?

KNardi — Tue, 03 Nov 2009 16:28:37 GMT

VSphere 4.0, VCenter 4.0, ESX hosts = 9 all using HA and DRS on one Cluster

My question is, Is it possible to deploy a PC workstation using LANdesk on the VIC?

Can we use a PXE boot?

AMD Opteron 2435 2.60 GHz Six Core VS Quad Core AMD Opteron(tm) 8384; 2.7GHz

fanzalone — Mon, 02 Nov 2009 21:31:57 GMT

Dear All,

I'm new in the vmware community.

I'm using vmware Enterprise 3.5 since 3 years.

I need to change two of my vmware hosts and I would like to know which benchmark score this two different machines have.

both machine are dual socket equiped.

Thanks a lot for your help.

What VMware product to use? Your dream setup high-powered personal workstation...

EricBryant — Fri, 30 Oct 2009 20:20:47 GMT

Hello All,

If you want to run multiple guest O/S's for your personal workstation, and want the absolute highest possible desktop performance ever (including graphics performance) and you had a big beefy host with plenty of CPU & RAM to put it on, would you run vSphere, ESXi, or Workstation 7 on it (all cost considerations aside) and if you chose Workstation, what base O/S would you use assuming you want maximum RAM available to guests?

Thanks!

Eric

VM suddenly running VERY slow

mattstewartcsd — Fri, 30 Oct 2009 10:31:20 GMT

Hi, I am running Virtual Servers on an ESX 4.0 host. The host has plenty of grunt and all servers are running ok,....apart from one. All of a sudden I can no longer RDP to to (although I can see it on the network and ping it). I can logon the the server using the console but it takes an age to do so and sits on the 'applying computer settings' window for minutes at a time. The server is Win 2K3 R2 x86 with 2 vProcessors, 4gb RAM, and plenty of disk space. It was running fine a few days ago.

Any ideas what can casue this?

vSphere vcpu scheduling imbalance

ltbraswell — Thu, 29 Oct 2009 19:16:11 GMT

My group has been doing some testing comparing virtual to bare-metal performance with a particularly high-compute utilizing application.

Platform: Intel E5540 (Nehalem) quad-core, dual-socket. 48GB of 1066MHz memory (properly balanced)

OS/guest OS: RHEL 4.7 (for both bare-metal and virtual configs)
Configuration: Hyper-threading off for both esx and bare-metal. Numa enabled and verified on both ESX and RHEL 4.7 (bare-metal)
ESX version: 4.0u0 build 164009
ESX host power.CpuPolicy is set to "static"

The application is a single-threaded modeling/simulation type program. It is not I/O intensive at all and basically each job generates a load average contribution of 1 (i.e. it pegs a cpu for each job). What we have been doing is running 1,2,4, and 8 simultaneous jobs and comparing the virtual to bare-metal physical performance with different combinations of VM vcpu counts (1,2,4 vcpu). Don't know a lot of details about the application other than when scaling from 1 to 8 jobs on bare-metal, the performance of any one job degrades somewhat so it seems to be memory constrained (not sure if the is sensitive to last level cache size or raw RAM performance). The "model" size itself for each job is 1.5GB.

The observation is that when simultaneously running 2 and 4 of these applications, each in its own 2vcpu virtual (so 2 and 4 virtuals respectively but all running on one 2-socket physical). More often that not for the 2 job case, we will see the worlds representing these active vcpus both running on 1 socket (with the other socket basically idle). For the 4 job case (though not as frequently), we have seen the load split 3/1 across the 2 sockets. When we run 8 jobs we get an even distribution of active vcpu worlds across the 2 sockets (4 on each).

The question is: Why is it that when there are 2 active vcpus does ESX schedule these vcpus on one socket more often than not. Same thing for 4 active vcpus, lopsided scheduling 3/1, though not as frequently. The hypothesis is that this behavior is related to some power management function of the scheduler (i.e. it would rather load up one socket more fully so that the cores on the other socket have a chance to be freq scaled down or in the halted state and therefore save power).

Ordinarily for an application that is not sensitive to memory performance or last level cache size this might not be an issue but for our application it apparently is sensitive the load being spread across sockets.

We can workaround this lopsided scheduling by using affinity rules to pin virtuals to sockets and manually force the load to be spread but we were hoping for some way to get it to balance the load without using affinity rules. Note that our hosts are running with power.CpuPolicy static.

Any ideas?

Message was edited by: ltbraswell changed subject to remove the reference to version 4

ESXi 4.0 shows incorrect memory usage at host level

dasbacgl — Thu, 29 Oct 2009 17:33:56 GMT

Hey,

I've installed the Nehlem patch on my ESXi 4.0 boxes (current build #193498). The individual guest VM's now report their Active memory correctly, and Memory Usage is also correct, however at the host level it now appears that ESX is reporting Memory Usage (%) based on Consumed memory instead of Active Memory.

Has anyone else encountered this?

Thanks,

Garrick

Hi Guys.. my VMware get conrinuous reboot

kurva1979 — Tue, 27 Oct 2009 05:37:38 GMT

Hi Friends,

I am using vmware workstation and installed solaris 10, every thing look good for 1 month and later it get loop reboop( it get reboot continuour at GRUB options after 10 seconds) you can find the attachment .. pls can any one help me how to resolve this issue.. as it is very important for me pls guys.. help me to come out from this issue.. tks in Advance...

opinions on the 10gb network cards

cxo — Mon, 26 Oct 2009 19:19:49 GMT

Would like to know people's opinions on the 10gb network cards. In particular the "Intel 10 Gigabit AF DA Dual Port Server Adapter" and the "Chelsio Communications S320E: 10 Gbe Nework Adapter".
Just looking for horror stories.

I keep getting an error message "Error Caused by File"

kevinadavis — Sat, 24 Oct 2009 08:27:47 GMT

I am brand new to the virtual world and just recently installed blade servers and VMWare. I am running VSphere 4.0. I created my first VM Template and when I try and create a new VM from that template I get an error message "Error Caused by File" and the VM creation terminates. I created a 600 GB volume on my SAN to place the VMs I create so space should not be an issues since I only have my virtual center VM currently in that volume. Any help will be greatly appreciated.

Thank You!

Kevin

Resource Pools

Mr G Grant — Fri, 23 Oct 2009 10:11:55 GMT

We are currently hosting 417 virtual machines on a cluster containing 11 servers. Should i be looking at creating resource pools now as some machines are performing at bit worse now.

Any advise is much appreciated.

A Newbie needs help in a big way....HELP!!!

Dinube — Thu, 22 Oct 2009 21:19:50 GMT

Aloha:

I just installed VMware-Fusion-2.0.6-196839.dmg on My MacBook Pro running OS 10.5.8. It seems to work well.

However I am having problems installing a Hanger 9 simulator. It's on 4 separate disks. The first disk goes on ok asked for disk 2, but does not eject disk one.

I get an error panel which is attached.

Any ideas??

Thanks
Dan Page

VA deployment degradation on vSphere4

Lexus16 — Thu, 22 Oct 2009 11:08:21 GMT

Hello,

We have high performance degradation when our Appliance is deployed to ESX server via Virtual Center. When the appliance is deployed to ESX directly we don’t have such problem.
We are using the following script:

curl -u username:NfelMC2Ub -k -T upload https://10.250.148.27/folder/LR1_11133_ESXAppliance/upload.dat?dcPath=Center&dsName=Storage1

where 10.250.148.27 is the address of Virtual Center. If we specify the address of ESX instead of Virtual Center, everything works fast.
BTW such problem appeared on vSphere 4, on previous versions everything was fine.
Could you please answer:

What could be the reason of this degradation?
How can we tune up the speed of deployment via Virtual Center?
If we have correct credentials on Virtual Center, is it possible to deploy directly to the ESX registered on it?

Thank you

Performance issue when deleting snapshot

chrisAMS — Tue, 20 Oct 2009 13:52:38 GMT

Hi,

I used a tool to replicate my servers from one site to another. The problem is it happened often that the snapshot is huge say 10-20Go. So when the delete process begins, the ESX becomes slow

and sometimes completely unaivailable. I would like to know if there is a way to increase performance by for instance adding memory, or if there is a way to limit resources when comiting a snapshot

to this specific task so the delay will increase but keep good performances on the ESX.

Thanks a lot for your help

Listing hosts from a failed VM ESX 3.5 server crash

Just_Bob — Tue, 20 Oct 2009 12:34:16 GMT

Had a VM server crash the other day in our cluster. All of the hosts migrated to other servers with no issues. Is there a way that we could tell what host were on the server at the time it crashed? Would make life easier for our operations unit.B-)

vmware server performance benchmarks...

eggie — Mon, 19 Oct 2009 20:52:27 GMT

Hello - I am looking for independent studies showing benchmarks/performance of vmware products on different hardware platforms (dell, ibm, hp, etc.). I have only been able to find studies produced by the vendor (ie. ibm, hp, etc.) and am seeking something more objective.

Thank you.

ESX 3.5 VM Perfomance Charting

vmproteau — Mon, 19 Oct 2009 16:53:45 GMT

ESX 3.5 (build 163429) Virtual Center 2.5 (build 147633)

In our environment, we have a predominance of WIndows servers. When asked for VM performance data, we typically recommend the client perfom OS level performance monitoring for detailed information. There are an increasing amount of requests where Virtual Center metrics and performance charts suffice and I see us getting more requests to provide this level of perfomance data and charting.

When charting in Virtual Center, I notice that realtime data has more metric options than day, week, month, year and custom. Particularly in disk, outside of real-time, I only have a disk usage metric.

First, is what I descrie what I should expect to see?
Is there a configuration option (or detail option) I can set to allow those metrics to be chartable in the other time frames?
If there is, is it wise to make that change (will it have potential negative effects)?

Any information about performance charting for VMs withing Virtual Center would be helpful.

reinstalling after time machine back up

phydaux — Sat, 17 Oct 2009 10:16:56 GMT

Hi there,

I'm about to replace the hard drive in my mac book, I plan to reinstall everything
from time machine.

Has anybody done this? are there any special issues with VMware or XP, will all my XP
data be OK or should I reinstall it separately?

Thanks in advance.

PS I’m running Mac 10.5.8 VM ware
2.00 Windows XP.

IBM LS22 server and Windows 2008 Terminal Server CPU Maxing Issue

Nemus — Fri, 16 Oct 2009 16:31:33 GMT

I am running VM4 Sphere with a guest OS of Windows 2008 Terminal server on an IBM LS22 with 16 gigs of ram and 2 cpu's after the fith person logs in to the Win 2008 Terminal Server the cpu hits 100% on the virtual machine.

My assuption is that its an issue with Windows 2008 TS ,but thought I would do a post here to see if anyone could shed some light on the issue.

The system runs to a crawl after this happens and I haven't been able to run any other virutal machines.

This server is Runing VM V4 sphere

I've been runing V 3 with out an issue, but the new Sphere V4 seems to be buggy.

Any suggestions about what could be causing this issue would be helpful.

So to recap my questions are:

Is this a vmware issue or just an Guest OS issue?

If it is a vmware issue what could be causing the guest OS to spike to 100% on both cpus?

Thank You

2 vCPUs Hardware Interupts

Wadebum — Thu, 15 Oct 2009 21:24:08 GMT

We have decided to move our Oracle Forms servers to VMs. We built a 2 vCPU with 4GB of Ram Windows 2003 SP2 VM on a new ESX 3.5 host using its own LUN on a EVA8100. Our DBAs started to do some forms compiling on the VM and noticed that it was taking around 4 times longer on the VM than it does on physical servers. They wanted me to add more CPUs so I added 2 bringing the server up to 4 total. I watched the performance as they reran the forms compiling and noticed that the actual process doing the compiling was only using about 25% of the CPU (my thinking is that the process does not take advantage of multiple cores). I put the server back down to 2 vCPUs and after rerunning the compile it was using about 50% CPU (adding to my suspicion that it does not take advantage of multiple cores).

Since it was a new Oracle install we thought maybe there is something diffrent between the physical server and the VM, so I did a test with a script that gathers all the email addresses in our Active Directory and puts them in a file. I added a function to the script to put a time stamp at the top and bottom of the file so I could compare between the physical and virtual server. The physical server finished the script in around 6.5 min but the virtual server took around 16-17 min to finish the AD script.

Since we were about to upgrade to ESX 4 we decided not to spend much more time figuring this out. Our upgrade to ESX 4 is now complete and I have upgraded the VMware Tools and Hardware of the VM to the latest versions. It still takes around 16-17min to run the AD script.

I then built a brand new VM with 1 vCPU and 1GB of Ram in the same LUN as the origional server and installed Windows 2003 SP2 onto it and installed the VMware tools. After that I ran the AD script 5 times and it finished on average at just under 5 minutes. I then installed all available WIndows Updates and ran the AD script again with no significant time increase.

Next I shutdown the VM and added a vCPU. After finding the new hardware I ran the AD script 5 more times with an average around 17min. I watched the server with Process Explorer and the Hardware Interupts are running around 20% the entire time the script is running. I used KernView to see what was going on in the Kernal of Windows and I get processor, ntkrnlpa, hal, tcpip, win32k, e1000325, ndis, afd, fltmgr, and ntfs as the top 10 items.

Using Process Explorer on the Physical Server shows the Hardware Interupts around 2% at the highest.

I have gone so far as to make this the only VM running on an HP BL465 G5 with 2 Quad Core AMD 2.7GHz processors and 32GB of ram and setting up CPU and Memory reservations to what the VM has but it does not perform any better. Does anyone have any idea how to get better performance out of a VM with multiple vCPUs? Or why the Hardware Interrupts are so high?

Centos 5 VM 100% CPU usage

godpickny — Tue, 13 Oct 2009 14:15:44 GMT

Hi, I've installed ESXi at a few sites still getting a feel for ESXi. Its been a God send in large part, but I'm having some serious issues with CPU usage. At first I configured a Centos VM with a single cpu. This VM is running server software for about 30 clients, evrything was fine. But then I noticed that the server started moving slow, a check with system monitor showed the cpu at 100% usage. A reboot and things would be running fine again until a couple hours later when the same result occurs...100% cpu usage.

Even at nights when no one is logged on the cpu is still 100%. Adding a second cpu only helps slightly, but after a few days both cpus are at 100% usage. Is there something I missed in the config or something else I need to do. This same configuration Centos5 running on a stand alone pc works perfectly but once its virtualized these issues arise.

I'm running ESXi on IBM Server with quad zeon, 6GB RAM, RAID5

Centos5.2 VM configured with 3GB RAM and 2 vCPUs

I've seache the net for answers but none that applies to Centos can someone help me?

Horrible performance after long inactivity

jkounis — Sat, 10 Oct 2009 20:48:47 GMT

I am running a Windows XP (32-bit) guest on an Ubuntu 9.04 (64-bit) host using VMPlayer 2.5.3. Performance was marginal until I set "sched.mem.pshare.enable" and "mainMem.useNamedFile" to FALSE. Now performance is very good with one major exception:

If I leave the guest running overnight, when I start using it again n the morning, the performance is absolutely horrible until I reboot the guest. This morning, from clicking "Start" to clicking "Turn off this computer" was 45 sec. It took about 15 min to reboot the guest.

Is there a memory leak in VMPlayer or some other issue that causes performance to degrade over time? Is there an issue with suspend on the guest that causes somethig similar (maybe WIndows XP Power Saving or Suspend features)?

The host has 6GB of memory. I have 2GB allocated to the guest. The pertinetn information is below:

memsize = "2048"
MemAllowAutoScaleDown = "FALSE"
sched.mem.pshare.enable = "FALSE"
mainMem.useNamedFile = "FALSE"

How to handle Vendor\Application requirements (physical) in VMWare

gbattiston — Wed, 07 Oct 2009 15:02:34 GMT

We have thoroghly embraced virtualization in our organization and it seemed to be going smoothly until the Application folks start to get involved. We have had many App owners refuse to move to virtualized servers unless we give in to what they say are "Application" or vendor requirements. For example i am being asked to build an app server and i would go about it using our standard deploy of a guest with 1vCPU and 1GB RAM. The application folks are quoting vendor documentation saying that the server must have a minimum of 2 CPUs and 8GB RAM. If i were to give in to all the different vendor \ application requirements then my host would be full after only 10 VMs!!

is there a tool or some sort of logic that would help in breaking down application requirements, for a physical server, and translating that into actual or needed requirements in a virtual server? or is it simply that if the server needs 4 GB RAM in physical, then it needs it in virtual also?

Per-VM I/O Monitoring On NFS Datastores

deltajoka — Wed, 07 Oct 2009 08:53:17 GMT

I realize that this issue has been up on the forums before, but no real alternative solution has been posted. So the purpose of this post is to hopefully stir up some discussion.

We have recently migrated from local VMFS datastores to NFS datastores on a IBM System Storage N3300 Model A20 filer (basically a rebranded NetApp). It is no longer possible to get statistics of per-VM I/O in vCenter or by using esxtop on the ESX hosts. We have filed a Support Case to VMware about this, but there is apparently no solution on the VMware side of things. VMware just refers to monitoring this on the NFS filer side. On the IBM filer, we can only monitor I/O per volume (and having a volume per VM is just not feasible). There is no way (that we know of) to monitor I/O per file on the NFS filer.

We think this is a huge drawback for NFS datastores (which is quite popular). Quite frankly, we are surprised that VMWare don't seem to be interesting in implementing this.

Has anyone managed to implement an alternative solution to monitor this? We are currently trying to build a solution that promiscuously captures NFS traffic on the storage network with tools like tshark/wireshark. But progress is slow due to the hassle of tracking back filehandlers to specific files and directories.

Any constructive input is most welcomed!

Regards,

Johan Karlsson

.

Scissor — Tue, 06 Oct 2009 05:50:28 GMT

esxtop reports wrong zero values for CPU and Memory (Service Console)

Timo1970 — Mon, 05 Oct 2009 16:35:33 GMT

Hi everybody!

We are working on a BMC Performance Manager (PATROL) Integration based on the esxtop command in batch mode.

The integration works fine yet but esxtop reports zero values for several service console CPU and memory values but

they aren't zero at all as esxtop in interactive mode shows correctly?

Does anybody know about this problem.

It is ESX Server 3.5.0 Update 4 on IBM hardware ...

Thanks in advance and kind regards ...

Timo

Any Performance Advantages Using Multiple Virtual Hard Disks? (eg, SQL Server)

DonJuanJovi — Wed, 30 Sep 2009 18:52:08 GMT

Hi Everyone,

First post here (long time listener, first time caller, lol)...

I have always wondered, is there any performance gains to using multiple disks in a VM? A specific example would be a virtualized SQL server - it's always said you should seperate the database and log files (to keep them contiguous). Or to be more general, the swap file of any Windows OS should be placed on 2nd drive if possible.

Does this apply to VM's as well? It's not really a disk to begin with, and the virtual disks sit on a group of shared physical disks, so.....

If anyone has some insight or, better yet, can point me towards any document or benchmark concerning this, that would be great!

STAF failing to start in webserver and fileserver systems

qc4vmware — Tue, 29 Sep 2009 23:01:41 GMT

I just moved a working tile to a different subnet and had to re-ip all of the systems. I don't know if this could possibly have anything to do with staf also not automatically running on boot for the webserver and fileserver systems. When I reboot each system I have to now manually execute the /etc/init.d/start-staf.sh script on both of those nodes. Oddly the database system and all windows system seem to still be working fine. Anyone have an idea on what might have went wrong? I am stumped and it seems like a pretty straight forward thing. Can you point me to a log that might shed some insight?

Paul

Insatllation VMmark and pre requistions

creddya — Wed, 23 Sep 2009 05:06:28 GMT

Hi every one,

Can any body tell me where i have to install this in ESX Server or any physical machine i can use to create the set up

And tell me setup procedure..

Waiting for ur reply..

VMware ESX 3.02, c-Class Enclosures & HP EVA 8x00

VCP210708 — Mon, 21 Sep 2009 19:32:07 GMT

Hello,

Im hoping someone will be able to assist me with regards to read and write latency.

I have been doing some baseline storage reports for the EXA's that host our VMware Virtual Desktop Infrastructure.

The setup I have is HP c-Class enclosures running VMware. 2 x 8 Server clusters per enclosure. Connectivity and Server hardware are BL485 G5 Servers - HBA - mds9124e internal - mds9124e external - (NPV mode) - Core Switch - Storage Port (EVA) - EVA 8100.

Load is evenly distributed across all storage processor ports which the Controller Status Performance Object is showing. 1 controller is working slightly more as the BFS volumes all go through Port 1 on Controller A due to the installation LUN must be the lowest-numbered of the LUNs available to the physical host.

Do we have any figures as to limitations on the following EVAPerf counters?

Host Port Statistics Performance Object
Read Req/s & Write Req/s
Read MB/s & Write MB/s
Read Latency & Write Latency
Av Queue Depth

Queue Depth Readings are around 1-2 which looks to be fine, host port Stats look to be fine with utilization on the paired controllers below 50%

The issue where I need assistance is the Average Read Latency & Average Write Latency, I understand that Aa average Read Latency < 15ms and Average Write Latency < 5 ms should be acceptable for most workloads, does this also cound for VDI's?

I am seeing the readings below, read latency is fine, however the write latency is above the 5 (ms) recomendation, now this recomendation isnt set in stone and so if there are any concrete acceptable / not acceptable figures out there supplied by HP or VMware I would appreashiate a benchmark figure.

Totals	Average Read Latency (ms)	Average Write Latency (ms)
DATA_DG1	9.117820069	8.909515571
DATA_DG2	9.11710754	8.340271941

Kind Regrads

Deano

vSphere Broadcom performance issue

Astro — Wed, 09 Sep 2009 13:20:14 GMT

Hi all

My configuration: 2 x HPDL 380 G5, vSphere Standard, vCenter Foundation, MSA1000

Both servers are equipped with 1 HP NC373i Multifunction Gigabit Server Adapter (embedded - 2 ports) and 1 HP NC360T (added - 2 ports). The first is identify by ESX as Intel Corporation 82571EB Gigabit
Ethernet Controller and the second as Broadcom Corporation Broadcom
NetXtreme II BCM5708 1000Base-T. The NIC are connected to the same Gb switch.

When I copy a test file (1GB) from a physical server to a VM (Windos 2003) that is "binded" to the Intel NIC, performance are 40-42 MB/s. The same test from the same server to other VM (Windows 2003, XP, Linux) performs 20-22 MB/s.

Maybe a IRQ or a driver problem ?

cat /proc/vmware/interrupts (Broadcom are vmnic0 and vmnic1)

Vector	PCPU 0	PCPU 1	PCPU 2	PCPU 3	PCPU 4	PCPU 5	PCPU 6	PCPU 7
0x21:	0	0	0	0	0	0	0	0	VMK ACPI Interrupt
0x22:	1425133	1310196	980490	919533	742309	764753	976164	955231	VMK cciss0
0x29:	120	0	0	0	0	0	0	0	COS irq 1 (ISA edge)
0x31:	3	0	0	0	0	0	0	0	<COS irq 3 (ISA edge)>
0x39:	4	0	0	0	0	0	0	0	<COS irq 4 (ISA edge)>
0x41:	1	0	0	0	0	0	0	0	<COS irq 6 (ISA edge)>
0x49:	1	0	0	0	0	0	0	0	COS irq 8 (ISA edge)
0x51:	0	0	0	0	0	0	0	0	<COS irq 11 (ISA edge)>
0x59:	157	0	0	0	0	0	0	0	COS irq 12 (ISA edge)
0x61:	0	0	0	0	0	0	0	0	<COS irq 13 (ISA edge)>
0x69:	1275	5883	0	1020	1	0	4080	0	<COS irq 14 (ISA edge)>, VMK libata
0x71:	0	0	0	0	1	0	0	0	<COS irq 15 (ISA edge)>, VMK libata
0x79:	0	0	0	0	0	0	0	0	<COS irq 16 (PCI level)>, VMK uhci_hcd:usb2, VMK ehci_hcd:usb1
0x81:	0	0	0	0	0	0	0	0	<COS irq 17 (PCI level)>, VMK uhci_hcd:usb3
0x89:	6621097	3820460	21339423	22894321	15919549	14392462	14777265	15413282	<COS irq 18 (PCI level)>, VMK qla2xxx, VMK uhci_hcd:usb4
0x91:	257081	123218	249195	227139	159631	332055	172800	166005	<COS irq 19 (PCI level)>, VMK qla2xxx, VMK uhci_hcd:usb5
0x99:	0	0	0	0	0	0	0	0	<COS irq 22 (PCI level)>
0xa1:	0	0	0	0	0	0	0	0	<COS irq 20 (PCI level)>
0xa9:	76	0	0	0	0	0	0	0	<COS irq 21 (PCI level)>, VMK uhci_hcd:usb6
0xb1:	0	0	0	0	434940540	0	0	0	VMK vmnic0
0xb9:	0	0	0	0	3791755	0	0	0	VMK vmnic1
0xc1:	1683702	4355657	916470	1407856	520711	3391248	1266330	3719433	VMK vmnic2
0xc9:	154275	214610	102000	55080	58398	52275	68595	58650	VMK vmnic3
0xdf:	610585043	679585878	689560293	691393738	692551003	696278389	691877800	694125796	VMK timer
0xe1:	70913	310979	1297883	991128	778091	837844	817038	768270	VMK monitor
0xe9:	142000793	110612476	259330606	281321494	136070039	188644905	193187899	197254694	VMK resched
0xf1:	42010	278812	323655	291608	414655	398176	346804	311322	VMK tlb
0xf9:	30107670	0	0	0	0	0	0	0	VMK noop
0xfc:	0	0	0	0	0	0	0	0	VMK thermal
0xfd:	0	0	0	0	0	0	0	0	VMK lint1
0xfe:	0	0	0	0	0	0	0	0	VMK error
0xff:	0	0	0	0	0	0	0	0	VMK spurious

Many thanks and regards

Max

SAP on VMware: How to interpret CPU Time reported by Workload Monitor (ST03)?

Tim Wise — Tue, 08 Sep 2009 17:00:47 GMT

SAP's Workload Monitor reports transaction CPU Time. This is amount of CPU seconds consumed by different types of transactions. There's a total for all transactions and an average per transaction.

I'm aware of the Time Keeping in VMs issue and am wondering whether the SAP reported CPU Time is affected by this issue.

Or has SAP's collector been modified to collect the CPU time directly from the host (rather than from the VM)?

Any info helps. Thanks.

Tim

Virtual Center Server -- Performance issue - VPXD.exe and java.exe are taking high CPU usage ....50 or 56% like this...Please advice on this ...

Kitu045 — Mon, 31 Aug 2009 21:09:49 GMT

Pradeep K Yadav
Pradeep045@gmail.com
VMware & VDI administrator

Performance testing VMware thinapp client server application

Ebyjacob — Tue, 25 Aug 2009 20:09:37 GMT

Hello All,

We have an application which is developed with vmware thinapp. The application startsup when an Exe is invoked. Which is the best performance testing tool which can be used for performance testing. We have tried JMeter, Rational Performance tester, Loadrunner but none of these tools are recording the actions performed. The protocol which is used is FTP. It would be great if anyone could give a possible solution for this.

Many Thanks

Eby

Two VMs Much Worse Than One - IOWAIT?

tkcook — Mon, 24 Aug 2009 21:21:25 GMT

I have VMWare Server 2 running on Ubuntu Jaunty desktop. The host is a core 2 quad with 2GB of RAM and two disks in RAID-0.

The guests are one Windows Server 2003 with 1 CPU, 700MB RAM and two SCSI disks of about 20GB each, and one Windows XP Pro with 1 CPU, 512MB RAM and one SCSI disk of about 20GB. There is not a great deal else happening on the box - it should be able to hold both memory images in physical RAM, I think, especially since it shows actual usage considerably less than what is allocated.

When I run either of these alone, everything goes very well; the guests are snappy and the web access page shows that the virtual CPU is running at around 1.6GHz - easily enough for my needs.

But when I try to run both of them simultaneously, the whole system grinds nearly to a halt. The two virtual CPUs between them manage about 250MHz (as shown by the web access page) either even split or about 200 to one and about 30 to the other. Needless to say, they are both unusably slow.

Digging around, the host's top shows that the CPUs are spending around 80-90% in iowait - with only one VM running this drops to about 3%. iotop shows that there are frequently up to about forty vmware-vmx threads all trying to read about 100KB/s, which probably explains things - the disks are madly dashing from place to place to fill all the requests. The overall transfer rate is pretty tame - about 2-4MB/s, well below the capabilities of the disks in sustained transfers.

Both virtual machines have:

sched.mem.pshare.enable = "FALSE"
mainMem.useNamedFile = "FALSE"

in their VMX files.

I've Googled for several hours and have found a number of people describing similar problems, but none with many solid suggestions and nothing that has made much difference (converting one VM from IDE to SCSI disks helped a little, but not significantly).

Can anyone suggest how to get these two to play nice together?

wiutyr

mmd7777 — Mon, 24 Aug 2009 14:35:20 GMT

fuy9uyer98

Delay in alarm triggering???

johnmack — Mon, 24 Aug 2009 08:49:43 GMT

Hi,

I am trying to create a script that will run as soon as the alarm is triggered. The issue I am having is that when the the CPU usage goes above the trigger the alarm is taking 2 hours to react?

I am asking if anyone has come accross this delay in reaction using alarms. (ie This is being carried out in a test environment where when I run the stress test it spikes to 100% CPU usage for 3 mins and the alarm takes about 2 hours on average to actually activate?????)

Is there a way to reconfigure the alarm to be more reactive and responsive to CPU usage spikes??

Any help appreciated

John

Poor VM Performance when running CentOS 5.3 on ESXi 4

joshualim — Fri, 21 Aug 2009 03:35:24 GMT

Hi there,

I'd build a VM which runs on CentOS 5.3 32bit, and had only WHM/cPanel installed. I had only a few sites on the VM, however, when i login to SSH of the VM and monitor 'top' , httpd , lfd, php sessions are using high CPU percentage. When i use 'dso' as the php handler, each httpd session will use 7% CPU, but when i change to CGI / suPHP / FCGI, httpd cpu usage drops to 0.3 - 0.5 % per session. However, these percentage are still high.

My sites are not really high traffic and the VM has been assigned 4 cores (Core2Quad Q9400) and 2 GB RAM allocated to it (with a server total of 4GB RAM) .

Is there anyway to config / optimize the settings of ESXi ? As i assume its ESXi problem as I had been using CentOS 5.x with WHM on many servers before trying VMware ESXi and none had this problem .

Do appreciate if anyone can help.

Thanks in advance.

Cheers.

Josh

Performance issue with DOS-based application on Citrix

CaptOchs — Wed, 19 Aug 2009 13:54:54 GMT

I'm having some performance issues with a large DOS based ERP system ever since migrating to VMWare. The ESX server is running 3.5 on a Dell 2950 G3. The guest O/S is Windows 2003 SP2. The VM has 2 CPUs and 2GBs of RAM assigned to it. The users access the application VIA Citrix Presentation Server 4. Certain functions (like running reports) are taking a long time. Some reports used to take about 8 minutes now take 35 minuts. I've ran some tests on a matching test environment and this is what I've come up with:

Sales Report, Key is Configuration = test1, test2

Single CPU configuration = 1:45, 1:40

Dual CPU configuration = 2:00, 2:30

Back to single CPU w/ Uniprocessor HAL = 1:20, 1:30

Single CPU w/ Multiprocessor HAL = 1:20, 1:25

Increased from 2GB to 4GB RAM = 1:20, 1:20.

This sales report used to take about 25 to 30 seconds on the physical hardware. I'm forced to run at Dual CPU because a single CPU can't handle the user load. Citrix Resource Monitor indicates my Interrupts percent on the CPU is high. I know DOS applications are interrupt-intensive. I've tried various settings in my PIF file and even a third party product called TameDOS. Nothing seems to help. Anyone have any ideas?

VC performance counters dropped hourly

Thibault — Mon, 17 Aug 2009 15:26:43 GMT

It seems that on an hourly basis my VC server drops performance counters for 10 minutes for each ESX server in the environment. Each ESX server on a "different" hourly schedule seems to be missing 10 min worth of counters (not every hour but most). This missing data seems to bring the average down for all counters overall causing a skew in averages. My VC server is a VM running 2.5.0 update 147633, my Oracle DB is on a remote physical server. It does not look like VC VM is being taxed and am told the DB server is not real busy either. I am logging data in Info (normal logging) mode with the normal default Statistics Intervals of 5 min for 1 day, 30 min for 1 week, 2 hours for 1 month and 1 day for 1 year. Does anyone have any ideas what may be wrong? Thanks

2. What is the default permission when a partition is formatted with NTFS

bazlina — Wed, 05 Aug 2009 15:20:49 GMT

1. What is the default permission when a partition is formatted with NTFS

2. What is the best way to secure files and folders that i share on NTFS partitions?

copy from hypervisor to nfs very slow

c20let — Tue, 04 Aug 2009 14:23:15 GMT

Hi,

I'm try to copy the vmdk file from the esxi server to an nfs and this transfer rate is very slow approx 9Mbps. Is there a particular limit or parameter which I can set this to 100Mbps. Also if I try to download from the datastore to my a computer again transfer rate is approx 9Mbps.

Anyone got ideas

cheers

Daniel

Drummonds vs. Crosby: The Virtualization Debate

yorkie — Thu, 30 Jul 2009 21:22:20 GMT

The debate at Burton Catalyst was awesome!

Chris Wolf did a great job of chairing a strong debate: both Scott and Simon did well, delivering startlingly different approaches which will be no shock to anyone familiar with Simon Crosby's conjecture approach versus Scott Drummonds' data approach.

I did some googling to check into the backgrounds of both and found some startling results!

Read more analysis and find the video at ViewYonder

Lenny slow

FilbertL — Wed, 29 Jul 2009 16:56:52 GMT

Dear,

We're running 2 ESX servers 3.5 with several Debian VM servers.
All of them are running Etch 4.x with kernel 2.6.18 and VMware tools and 900 MB memory (no GUI)
Now I've upgraded a test-server to Lenny 5.0.2 running kernel 2.6.26 also with VMware tools and 900 MB memory and no GUI.

It seems to me that the Lenny runs significantly slower than Etch. A task that takes 6 seconds on Etch, takes 11 seconds on Lenny.

Booting with the "old" kernel 2.6.18 does not help.

Who else is experiencing the same? Any solution for this?

Thx F.

Linux virtual machine gaining time

jamcm — Fri, 24 Jul 2009 22:11:59 GMT

I have a red hat linux 4 update 8 vm that is gaining time at a rate of about a second per hour. The VM runs on a physical blade esx server running esx 3.5 update 4 and it is the only vm on the host.

vmware tools time sync is turned off, we use ntp syncing to a physical piece of hardware. We followed a suggested kb and set the clock setting in grub.conf

this did not seem to help.

we also set the ntp.conf to have tinker panic 0

Anyone ever run into this?

ESX 4 and iSCSI Speed

JMBCalgary — Fri, 24 Jul 2009 16:42:14 GMT

Hey Guys,

We are doing some testing with iSCSI and finding some results I am unsure of. Here is our Setup-

ESXi host:
6 network ports - 2 on board, 4 on Card
2 Ports are setup for Managment and Vmotion
4 Ports are for iSCSI and VM Port Groups

The 4 ports are setup in a team and I have the Load Balance set to Route based on IP hash
We have the 4 associated ports setup on our switch as a trunk with flow control enabled

On our iSCSI storage we have 2 1GB ports setup in a bond
and the 2 ports are setup on the switch as a trunk with flow control

I have build one VM on this host against the iSCSI array and done testing and get a little over 100 MB/s against the disk

If I take and split 2 of the 4 ports out on the host and use MPIO on the host I can get 186 MB/s but then that custs my available network speed for my VM's in half.

My long winded question is if we change the 4 Port nic out for a single 10 GB nic and use the same bonding on the Storage device will we see any better results without using MPIO on the ESX software iscsi initiator.

Jesse

Testing storage performance

aurimask — Tue, 21 Jul 2009 13:20:45 GMT

Hello,

I was testing ESX server storage performace, by using IOMeter inside virtual machine and default testing configuration. I got some results, but no results showing in esx performance monitor while testing. What i did wrong? Virtual machine is stored in local SAS disk. I get the result like that:

'Test Type	Test Description
0	diskas e
'Version
2006.07.27
'Time Stamp
2009-07-21 15:14:10:174
'Access specifications
'Access specification name	default assignment
Default	0
'size	% of size	% reads	% random	delay	burst	align	reply
2048	100	67	100	0	1	0	0
'End access specifications
'Results
'Target Type	Target Name	Access Specification Name	Managers	Workers	Disks	IOps	Read IOps	Write IOps	MBps	Read MBps	Write MBps
ALL	All	Default	1	1	1	239.686.869	160.480.912	79.205.956	0.468138	0.313439	0.154699
MANAGER	00DVSTEST	Default		1	1	239.686.869	160.480.912	79.205.956	0.468138	0.313439	0.154699
PROCESSOR	CPU 0
PROCESSOR	CPU 1
PROCESSOR	CPU 2
PROCESSOR	CPU 3
WORKER	Worker 1	Default			1	239.686.869	160.480.912	79.205.956	0.468138	0.313439	0.154699
DISK	E:Database					239.686.869	160.480.912	79.205.956	0.468138	0.313439	0.154699
'Time Stamp
2009-07-21 15:16:40:262

VCB 1.5U1 backup preformace on Fiberchannel Very slow 5mb/s

Don.G — Fri, 17 Jul 2009 05:17:22 GMT

we have 4gb/s fiber channel, Emc cx3-20 san and backing up from raid 1+0 8 disk array. all backups from this array and anyothers in our san only backup at 5mb/s using either full VM or FIle level VCB backups.

When i ran a disk benchmark i get 150mb/s on that raid.

How can i make backups faster? how do i even start to isolate this issue? what kind of preformance from a 1+0 should i recieve(15000rpm fc)? also on our sata raid (7200rmp 1TB drives) on the san its a 10disk raid 5 and we only get 25mb/s this seems drastically lower then it should be.

I welcome any help on this matter

Don

Limiting factors for iSCSI SW HBA / only 100MB in VM

alex123 — Thu, 16 Jul 2009 14:16:08 GMT

Hi everybody,

first i try to explain my problem.
We configured an iSCSI Target with SLES10.3. We can write (20GB) from another Server to that Target over NW with eg. 400MB/s. Using the iSCSI Target.

Now we configure our ESX Server to use this Target as HBA. Create a VM on an NFS Storage, create a hard disk on this Target. Now we repeat the writettest. We're not able to get more than 100MB/s.
Why? Can we change something? If its normal, which element decrease the speed?
I spent much time to find a solution for this problem. But i find4 nothing helpful.

HW Config:
ESX Server (ESX 3.5 build 153875):
HP DL 380 G8, 32 GB RAM, 8 x 2,833GHz, NW 2 x 10Gbit Ports with CX4.

Storageserver:
No Name Server / 8 x 500GB Raid 0 / NW 2 x 10Gbit Ports with CX4.

Thanks for helping, if you need more infos, write short info.

BusLogic Adapter Performance?

mahtoji — Thu, 16 Jul 2009 13:34:14 GMT

Hi,

When my FreeBSD VM writes to the bus logic HBA I suspect the interaction takes a long time. As a result the IO performance is low. Are there tools I can use to debug further. Both esxtop & vscsiStats do not seem to directly letting me peek in to the VMM.

Right now in the FreeBSD Bus Logic HBA driver I have instrumented (blue) the code as shown below. Towards the tail end of btexecuteccb I have -

bt->cur_outbox->action_code = BMBO_START;
if (trace_on_panic & 32)
microtime(&hbtv);
bt_outb(bt, COMMAND_REG, BOP_START_MBOX);
{color:#0000ff} if ((trace_on_panic & 32) && (xpt_path_comp(cam_dpath, ccb->ccb_h.path) == 0) && ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_OUT)) {
microtime(&hetv);
hbadelta = hbadelta + (hetv.tv_sec*1000000 + hetv.tv_usec) - (hbtv.tv_sec*1000000 + hbtv.tv_usec);
hbacount++;
}{color}
btnextoutbox(bt);

For a total of about 300,000 trips through this code I find about 14 seconds are spent in the entire function of which 10.7 seconds are spent in the about bt_outb alone! I suspect this is because of a very high emulation overhead of the BusLogic HBA inside the VMM. Any comments/pointers on how I could get more visibility in to the VMM or this issue? Pointers on how this could be reduced?

Regards,

how to use VMKPERF tool

knatesan — Thu, 16 Jul 2009 10:01:04 GMT

hi,

can someone help me how to use vmkperf tool.

what is the parameter i need to pass to capture cache missing and etc..

highly appreciate if i got any help document for VMKPERF.

Thanks,

kamal.

VCB slow performance backing up Novel VM

tpeter — Tue, 14 Jul 2009 16:49:48 GMT

When using VCB to backup the Novell virtual machine I get a failure message about not being able to queue, if I add the "-Q 0" option then I can get it to backup but it is very slow. Novell 5.1 sp8 guest installed on a ESX 3.5up4, VCB 1.5.0.2192, Virtual Center 2.5 update 4. The server is 18GB but takes over 10 hours to complete. If I do a windows server 60GB it will complete in 2.5 hours with no problems.All virtual machines are on 1 iscsi lun.

DNS and VMmark

tbjs — Mon, 13 Jul 2009 17:33:58 GMT

I am setting up VMmark for the first time, and I am having trouble understanding the dns setup. When configuring the mail server as a domain controller, the instructions tell you to enter the DNS name for the new domain, saying "this should be the FQDN in the format "mailserverN.yourcompany.com". So am I to assume this is my dns domain suffix? If so, why does only client0 follow this convention in the sample hosts file: client0.mailserver0.yourcompany.com, while all other are in the format "server.yourcompany.com" ? Maybe there is a difference in the concept of "domain" between active directory and dns?

FYI, I have completed the setup, but when I kick off the benchmark it fails trying to ping the server names (webserver0, mailserver0, etc) I am sure there is something wrong with my dns setup, so that is why I am trying to understand what I am looking at. Admittedly, I have not used Windows much, so AD and Exchange are a mystery to me.

Thanks,

Tom

Discrepancy between ESXTOP metrics and vSphereClient metrics.

knatesan — Mon, 13 Jul 2009 06:16:12 GMT

hi,

I am seeing some discrepancy between esxtop metrics and vcenter metrics. below is the metrics for the 1CSU vCON workload (HTOFF)

Physical CPU %Total Util Time = 55.65%

%Total Processor Time = 58.92%

vCenter Metics( Average used cpu) = 36.48%

Why this difference is here?

fyi, i have completely gone through the "understanding EXSTOP" article. but didnt help.

Below is the build i have used :

vSphere Client : ver 162856

ESX Host : ver 164009

Highly appreciate your time and help.

More eager to get the pin-point response.

Thanks,

kamal.

Newer desktop hardware vs older server hardware

relegated — Mon, 06 Jul 2009 05:27:00 GMT

I am currently running a Pentium D 915 with 2GB DDR2 and 80GB SATA drive and a 10/100 NIC all on a Gigabyte motherboard. I have been having problems, for example I get kernel panics when installing CentOS 5.3 http://communities.vmware.com/thread/208117 and can't get Ubuntu 8.04 server to install. I believe both of these are related to my processor/motherboard combo. With that said I am considering possibly switching to older server hardware which would be dual Intel Xeon 2.0GHz CPU - SL6EM - 2000DP/512L2/400/1.50V on a Thunder i7501 motherboard with registered ECC memory, not sure how much I can get for free and some 10k SCSI drives. My hope is that switching to this hardware would get rid of the problems I am having without a huge hit in performance but I don't know honestly. I am also considering the switch as the motherboard has dual gigabit NICs which would allow me to upgrade to ESXi 4.0.

So I just wanted to get some thoughts, is it worth the change?

Jade database apps / workloads

jpreou — Sun, 05 Jul 2009 23:45:42 GMT

Quick question, since a Google search and forum search hasn't turned up 'The Goods'.

Has anyone out there done P2Vs of jade database / application servers and, if so, seen any post-virtualization performance issues?

We have two engineers here in NZ that have recently done this for two different customers with different hardware bases (both using ESX3.5i) and both have seen degradation in the performance of their Jade systems. Multiple fiddling with 1vCPU vs 2vCPUs, affinity, reservations, shares, memory, reduction of load on the hosts, etc, etc, all the usual stuff, hasn't helped. If you review the ESX hosts it would appear that the hosts are not at all under load and the VMs in question are getting all the resources they need when they need it, and yet performance is still not as it was on the physical systems, even when those physical systems were lower spec'd.

I don't want to go into detailed troubleshooting here as it could take a loooong time and we are under pressure to find a resolution (short of going back physical which would cost significant $$$ now since the original servers have been re-deployed elsewhere). But if anyone out there has any experience or advice regarding Jade database systems, or can point me in the right direction to web-based resources I may have missed, then we would most certainly be grateful.

I'm also in discussion with another customer about virtualizing most of their infrastructure (15 or so VMs across three hosts and a SAN) and they have some Jade stuff too, so I'd like to be able to give that customer some confidence that we can VM those systems ok.

Cheers

Jeff

SAP MRP running on VMware 3.5 with parallel processing

Jimmya11 — Thu, 25 Jun 2009 23:12:54 GMT

SAP MRP recommends running parallel processing using 1/2 the available amount of CPU's available. How would I configure this in VMware 3.5, when the physical box has 24 CPu's, and runs 4 VM machines each with 2 CPU's? If I say '1', no parallel processing occurs, if I say more, what happens? Jimmya11

VM has no CPU utilisation but esxi 4 host shows high cpu usage

ThomasT74 — Thu, 25 Jun 2009 16:29:13 GMT

Hi,

I'm currently testing vmware esxi 4 inside of Workstation 6.5.2.

I created a new VM and installed W2k3 R2 SP3 on this Virtual machine. After the installation the Task manager schows low CPU usage ( < 10%) and ESXi host shows also low CPU usage.

After that I installed MS SQL Server 2005 with SP3 on this VM and after I started the SQL Server the CPU usage in Task manager shows low usage ( < 15 %) but the ESXi host shows a CPU usage above 80%. When I stop the SQL Server service the CPU usage on the ESXi host is back to normal, which is similar to the Task Manager usage.

Is there an explanation for this behaviour?

Memory Leak in WM or Linux?

ChrisWorks — Wed, 24 Jun 2009 08:48:48 GMT

Hello,

I have installed a Centos 5.3 on vmware esx server 3i, 3.5.5, 110271.

I have given it 2 cpu's and 4gb og memory. However it uses a lot of memory, but I cant find out on what?

This is how it looks after startup, 1,2gb

Tasks: 94 total, 1 running, 93 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 3896208k total, 1242996k used, 2653212k free, 23528k buffers
Swap: 5931000k total, 56k used, 5930944k free, 1117636k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
14390 gdm 17 0 31368 14m 7040 S 0.0 0.4 0:04.27 gdmgreeter
2181 root 15 0 13340 4792 1004 S 0.0 0.1 0:00.57 python
14369 root 17 0 14516 4584 3404 S 0.0 0.1 0:02.57 Xorg
2565 root 25 0 27460 4220 3580 S 0.0 0.1 0:03.03 gdm-rh-security
2367 haldaemo 18 0 5652 3748 1700 S 0.0 0.1 0:04.07 hald
2488 root 15 0 15528 2904 2344 S 0.0 0.1 0:00.15 gdm-binary
13656 root 15 0 9980 2852 2292 S 0.0 0.1 0:04.44 sshd

This while running a basic unzip command on a large file.

top - 10:26:46 up 19:04, 3 users, load average: 0.82, 0.27, 0.12
Tasks: 95 total, 2 running, 93 sleeping, 0 stopped, 0 zombie
Cpu(s): 8.1%us, 25.1%sy, 0.0%ni, 49.3%id, 17.2%wa, 0.2%hi, 0.2%si, 0.0%st
Mem: 3896208k total, 1857116k used, 2039092k free, 24364k buffers
Swap: 5931000k total, 56k used, 5930944k free, 1731900k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
15418 root 18 0 1844 604 452 R 65.1 0.0 0:37.41 unzip
15419 root 15 0 2196 996 796 R 1.3 0.0 0:00.56 top
16 root 10 -5 0 0 0 S 0.3 0.0 0:03.58 kblockd/1
1984 root 15 0 2340 940 764 S 0.3 0.0 2:54.63 vmware-guestd
13656 root 15 0 9980 2852 2292 S 0.3 0.1 0:04.52 sshd
1 root 15 0 2064 588 512 S 0.0 0.0 0:00.92 init
2 root RT -5 0 0 0 S 0.0 0.0 0:00.82 migration/0
3 root 34 19 0 0 0 S 0.0 0.0 0:00.10 ksoftirqd/0

After an unzip finished .. over 2gb. Running an unzip ate 1gb of memory

top - 10:27:52 up 19:05, 3 users, load average: 0.75, 0.37, 0.17
Tasks: 94 total, 1 running, 93 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.2%sy, 0.0%ni, 99.8%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 3896208k total, 2342816k used, 1553392k free, 24948k buffers
Swap: 5931000k total, 56k used, 5930944k free, 2217024k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
15419 root 15 0 2196 996 796 R 0.7 0.0 0:00.97 top
14147 root 15 0 9980 2832 2288 S 0.3 0.1 0:01.23 sshd
1 root 15 0 2064 588 512 S 0.0 0.0 0:00.92 init
2 root RT -5 0 0 0 S 0.0 0.0 0:00.82 migration/0
3 root 34 19 0 0 0 S 0.0 0.0 0:00.10 ksoftirqd/0

After running a tar -xvf finished it looks like this... 2.5gb

top - 10:29:53 up 19:07, 3 users, load average: 0.80, 0.48, 0.22
Tasks: 94 total, 1 running, 93 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.2%us, 0.2%sy, 0.0%ni, 99.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 3896208k total, 2545624k used, 1350584k free, 27016k buffers
Swap: 5931000k total, 56k used, 5930944k free, 2417200k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
15419 root 15 0 2196 996 796 R 0.3 0.0 0:01.67 top
1 root 15 0 2064 588 512 S 0.0 0.0 0:00.92 init
2 root RT -5 0 0 0 S 0.0 0.0 0:00.82 migration/0
3 root 39 19 0 0 0 S 0.0 0.0 0:00.10 ksoftirqd/0

after waiting 5 minutes from the tar job finished... it still uses 2.5gb of memory.

top - 10:35:17 up 19:13, 3 users, load average: 0.00, 0.15, 0.15
Tasks: 94 total, 1 running, 93 sleeping, 0 stopped, 0 zombie
Cpu(s): 1.2%us, 2.2%sy, 0.1%ni, 95.8%id, 0.5%wa, 0.1%hi, 0.1%si, 0.0%st
Mem: 3896208k total, 2545624k used, 1350584k free, 27420k buffers
Swap: 5931000k total, 56k used, 5930944k free, 2417200k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1 root 15 0 2064 588 512 S 0.0 0.0 0:00.92 init
2 root RT -5 0 0 0 S 0.0 0.0 0:00.83 migration/0
3 root 34 19 0 0 0 S 0.0 0.0 0:00.10 ksoftirqd/0
4 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/0
5 root RT -5 0 0 0 S 0.0 0.0 0:01.10 migration/1
6 root 34 19 0 0 0 S 0.0 0.0 0:00.10 ksoftirqd/1

I have just a simple centos 5.3 install. I have not installed anything myself. Eventually it reaches the 4gb max memory and everything slows to a halt ...

Why is it behaving like this???

Chris

Windows 2003 CPU MIBS - not gauge:32

jcarter — Tue, 23 Jun 2009 19:54:40 GMT

I'm trying to find an appropriate MIB to monitor CPU utilization on a Windows Server 2003 guest. We have physical servers (HP DL380-G4) where we monitor OID 1.3.6.1.4.1.232.11.2.3.1.1.2.0 from the guest OS. But after virtualizing it appears I need to look at a different MIB. The reason is that we load balance transactions to a farm of servers based on how not busy they are. This is done using a Radware App Director load balancing device that redirects web requests. unfortunately the only MIBs I have found so far are Gauge:32 mibs and the radware cannot use those counters. Any suggestions are appreciated.

Exporting Data for Clusters

JohnDi — Tue, 23 Jun 2009 12:12:06 GMT

Is there a way to export the performance numbers for a cluster out to a text file or spreadsheet file?

I would like to compare CPU & Memory data at the cluster level. We have several clusters, each for different CPU groups and I'd like to compare the performance of each cluster by graphing each cluster's CPU & Memory performance data on a single chart for CPU & Memory. I'd like to look at the data over the last 6 months.

I can go into each cluster and get performance data on the CPU and Memory separately, but it, a) just gives me that component, and; b) is only graphical. I'd like the numbers so I can create my own charts.

Any suggestions welcome.

Thanks,

John

%Ready is greater than 50. why?

knatesan — Thu, 18 Jun 2009 06:10:50 GMT

hi,

I am running 2 CSU vConsolidate workload in supermicro machine which has NHM processor(quadcore(0-15)) with SMT ON and Turbo mode ON.

Below is the performance chart:

% total util (Average) = 51 %

Power consumed = 320w

vCON geomean score = 3.43

Vcpu(1:idle:xxxx:idle(0-15)) % Ready = >50%

Question: my understanding is more %ready state will cause the CPU contention. Inthis case why %ready state is >50% ?

Please help me.

Let me know if you need any additional metrics.

Thanks,

kamal.

Large scale environment

curiousagain — Tue, 16 Jun 2009 19:13:20 GMT

Hi,

Can anyone give me an example of his Large scale environment (only working environment)?

I mean, a lot of VM's and total environment description (# of hosts, redundancy, SAN etc...)

I would also like to know what are the working details for any system parts (CPU usage, Memory usage etc...).

Thanks in advance.

Web service poor performance on 32 bit SLES11 in an ESX 3.5 env

durgarao — Tue, 16 Jun 2009 14:50:12 GMT

I have two environments: Both have same hardware specs.

1st Environment:

SLES10 SP2 32 bit is installed on a VM that is created on ESX 3.5 update5. On SLES10, Apache web server is running.

2nd Environment:

SLES11 32 bit is installed on a VM that is created on ESX 3.5 update5. On SLES11, Apache web server is running.

Note: On both environments, the VM is created with two CPUs/ 2GB RAM, 10GB HDD and 1GB network card. And on both environments, Apache is compiled from Apache 2.2.10 sources.

Using a HTTP test appliance (this is a WEB client, that generates HTTP_Get requests to the web server at a configurable rate), when HTTP_Get requests are done for a 4k object file, the WEB server on the 2nd environment gave poor performance (lesser than half) when compared that with the 1st environment.

With the 1st environment, on the top command output I have noticed that the CPUs are maxing at 7000 http_get transactions per second.

With the 2nd environment, on the top command output I have noticed that the CPUs are maxing at 3000 http_get transactions per second.

Also, an additional observation regarding hi (% CPU time spent on hardware interrupts) and si (% CPU time spent on software interrupts) CPU times from top command:

for 1st environment:

slowly increasing as the load increases to a max of 80%

for 2nd environment:

slowly increasing as the load increases to a max of 125%

On the second environment, ksoftirqd process in the 2nd case is consuming to a max of 18% (which should ideally not be more than 1% per cpu) and CPU is utilized at a quick pace.

Basically, It seems the problem lies with the ESX 3.5 support for SLES11 32 bit.

There is no specific install option for SLES11 on ESX 3.5. We tried both with "Suse Linux Enterprise Server" and with "Other Linux" install options from the drop down menu. Above observations stand in both install options.

Can anyone tell me what VM options on ESX 3.5 update 5 can improve SLES 11 32 bit performance?

If SLES 11 is not supported on ESX 3.5 update 5, when can we expect another update that supports SLES 11 32 bit in near future?

Thanks

Durgarao

variable IIS performance depending upon load on other ESX VMs

joeyCon — Thu, 11 Jun 2009 09:42:32 GMT

I am running a web server on an ESX box.

When I am running without any other VMs running on the ESX box, I get a throughput of 40 requests per second, without queues building up.

When I place other 4 VMs on the box (under heavy load, both CPU and network I/O) the performance of my web server deteriorates to 20 requests per second.

All VMs on single cores. All windows 2003, IIS 6

This happens if I set the affinity of the CPUs to set cores, or allow them to roam. If I set the affinity to a set core I also turn off hyperthreading.

There are serveral spare cores to allow for spare CPU capacity for virtual switches, etc...

Has anyone else had these issues? Am I missing something?

I'd be very greatful for any assistance.

Joe

Measuring the total cumulative time of all VMs and ESX

ssbangal — Tue, 09 Jun 2009 11:08:00 GMT

Hi,

In my experiment as I increase the number of Database VMs (SQL Server) that are running simultaneously on my host, beyond a certain point, the cumulative performance of all the Databases put together starts dropping. Looking at the performance graph that Virtual center generates, it is because of saturation in CPU utilization. In my experiments i am also capturing ESXTOP data every 2 sec.

What I am looking to do is analyze this ESX top data and try to capture the cumulative time of all the VMs spent and also time spent by the ESX server. As i increase the # of VMs, i am hoping to see increase in time spent by ESX server for scheduling, processing interrrupts etc.

Can you please let me know what fields i need to look up to measure the time consumed by the hypervisor?

Also, what is the best way to capture the cumulative time utilized by all VMs on the host?

FYI... I am not using any CPU affinity setting and no reservations. I am using the lastest version of vSphere.

Regards,

Sudhir

How do i get the latest , minimum and maximum value in real-time performance statistics using VI SDK

vmcts — Tue, 02 Jun 2009 13:52:38 GMT

Hi ,

I am working on real time performance statistics using vmware VI SDK. What I am getting using the vmware vi sdk is only the average value .Could any one of u let me know how do i get the latest, minimum and maximum value for CPU, Disk ,Memory etc using vmware vi sdk service. Please find below my code which i am using to get the real time data:

private void printEntityCounters(String entityType)

{

if (_ServiceContent == null)

{

_bConnection = false;

Connect();

}

Debug.Assert(_ServiceContent != null);

// Get Performance Manager

ManagedObjectReference pmRef = _ServiceContent.perfManager;

// Get supported perf counters in the system

PerfCounterInfo] cInfo = (PerfCounterInfo[)getObjectProperty(pmRef, "perfCounter");

// PerfInterval] cInfo = (PerfInterval[)getObjectProperty(pmRef, "historicalInterval");

ArrayList vmCpuCounters = new ArrayList();

int index = 0;

for (int i = 0; i < cInfo.Length; ++i)

{

if ("disk".Equals(cInfo.groupInfo.key))

{

vmCpuCounters.Add(cInfo);

index++;

}

Hashtable usageCounter = new Hashtable();

while (index > 0)

{

//int index = 0;

for (IEnumerator it = vmCpuCounters.GetEnumerator(); it.MoveNext(); )

{

PerfCounterInfo pcInfo = (PerfCounterInfo)it.Current;

Console.WriteLine(index + " - " + pcInfo.nameInfo.label);

PerfCounterInfo pcInfo1 = (PerfCounterInfo)vmCpuCounters--index;

usageCounter.Add((pcInfo1.key), pcInfo1);

//index--;

}

ManagedObjectReference vmmor = null;

_svcRef = new ManagedObjectReference();

_svcRef.type = "ServiceInstance";

_svcRef.Value = "ServiceInstance";

ArrayList morlist = GetDecendentMoRefs(null, entityType);

Boolean continueDataCol = true;

while (continueDataCol)

{

DateTime curTime = VS.CurrentTime(_svcRef);

DateTime beginTime = curTime.Subtract(new TimeSpan(0, 0, 20));

DateTime endTime = curTime;//.Subtract(new TimeSpan(1, 0, 0));//.Add(new TimeSpan(0, 5, 0));

if (morlist != null && morlist.Count > 0)

{

for (int j = 0; j < morlist.Count; j++)

{

vmmor = (ManagedObjectReference)((object[])morlist[j])[0];

ObjectContent] vmname = GetObjectProperties(vmmor, new string[ { "name" });

PerfMetricId[] aMetrics = VS.QueryAvailablePerfMetric(pmRef, vmmor, beginTime, true, endTime, true, 20, true);

ArrayList mMetrics = new ArrayList();

if (aMetrics != null && aMetrics.Length != 0)

{

for (int index1 = 0; index1 < aMetrics.Length; index1++)

{

if (usageCounter.ContainsKey(aMetricsindex1.counterId))

{

mMetrics.Add(aMetricsindex1);

}

else

{

Console.WriteLine("Virtual Machine " + vmname[0].propSet[0].val + " is not running");

}

FileStream filestream1 = new FileStream(PerformanceFile, FileMode.Append, FileAccess.Write, FileShare.Read);

StreamWriter SWW1 = new StreamWriter(filestream1);

SWW1.Write(SWW1.NewLine);

SWW1.Write("TimeStamp\t\tObject\tMeasurements\t\t\tUnits\t\tAverage\n");

SWW1.Write(SWW1.NewLine);

SWW1.Close();

filestream1.Close();

Console.WriteLine();

Console.WriteLine("Time\tObject\tMeasurements\t\tUnits\t\tAverage\n");

for (int count = 0; count < mMetrics.Count; count++)

{

PerfMetricId] metricIds = (new PerfMetricId[ { (PerfMetricId)mMetricscount });

PerfQuerySpec qSpec = new PerfQuerySpec();

qSpec.entity = vmmor;

qSpec.maxSample = 3;

qSpec.maxSampleSpecified = true;

qSpec.metricId = metricIds;

qSpec.intervalId = 20;

qSpec.intervalIdSpecified = true;

qSpec.startTime = beginTime;

qSpec.startTimeSpecified = true;

qSpec.endTime = endTime;

qSpec.endTimeSpecified = true;

PerfQuerySpec] qSpecs = new PerfQuerySpec[ { qSpec };

//Boolean continueDataCol = true;

//while (continueDataCol)

//{

PerfEntityMetricBase[] pValues = VS.QueryPerf(pmRef, qSpecs);

if (pValues.Length == 0)

{

Console.WriteLine("Virtual Machine " + vmname[0].propSet[0].val + " is not running");

}

else

{

//PerfSampleInfo[] infos = ((PerfEntityMetric)pValues[0]).sampleInfo;

//Console.WriteLine("Sample time range: " + infos[0].timestamp.TimeOfDay.ToString() + " read every" + infos[0].interval + " seconds");

for (int i = 0; i < pValues.Length; ++i)

{

PerfMetricSeries[] vals = ((PerfEntityMetric)pValues).value;

PerfSampleInfo[] infos = ((PerfEntityMetric)pValues).sampleInfo;

Console.WriteLine("Sample time range: " + infos.timestamp.TimeOfDay.ToString() + " read every" + infos.interval + " seconds");

for (int vi = 0; vi < vals.Length; vi++)

{

PerfCounterInfo pci = (PerfCounterInfo)usageCounter[valshttp://vi].id.counterId;

//if (pci != null)

//Console.WriteLine(pci.nameInfo.label + "-" + pci.unitInfo.summary);

if (valsvi.GetType().Name.Equals("PerfMetricIntSeries"))

{

PerfMetricIntSeries val = (PerfMetricIntSeries)valsvi;

long[] longs = val.value;

for (int k = 0; k < longs.Length; ++k)

{

Console.WriteLine(infos.timestamp.TimeOfDay.ToString()+ "\t"+ val.id.instance + "\t" + pci.nameInfo.label + "\t\t" + pci.unitInfo.summary + "\t\t" + longs[k] + " ");

FileStream filestream2 = new FileStream(PerformanceFile, FileMode.Append, FileAccess.Write, FileShare.Read);

StreamWriter SWW2 = new StreamWriter(filestream2);

//SWW2.Write(infos.timestamp.TimeOfDay.ToString());

SWW2.Write(infos.timestamp.TimeOfDay.ToString() + "\t\t"+ val.id.instance + "\t" + pci.nameInfo.label + "\t\t\t" + pci.unitInfo.summary + "\t\t" + longs[k] + " ");

SWW2.Write(SWW2.NewLine);

SWW2.Close();

filestream2.Close(); ;

}

Maximum no of Virtual machines

hrawat — Mon, 01 Jun 2009 08:30:25 GMT

I have a intel pentium dual core machine with 2 gb ram and 2 gige nic cards and i am trying to run a wireless protocol stack which can support maximum data rate of 512 Kbps on data channel and there is additional signaling load plus there are some management message. The stack runs on the VxWorks. I want to know how many VM we can start if all VM/stack start sending with the maximum data rate to peer entity ( outside the host machine) and what are the points that i should consider while calculating the no of concurrent VM that my machine can support.

rgds

Hsr

ESXI disk access slows down overnight

JaroF — Sat, 30 May 2009 19:52:57 GMT

I have actually posted this in the ESXI 3.5 forum, but I think this is probably a better place for it.

I have a very strange problem. I am running 8 virtual machines on VMware ESX Server 3i 3.5.0 build-158874 - it is an Intel Quad Core 8GB RAM machine. I am using a Perc 6 raid controller with 256Mb Cache and BBU. It was running for some time (three months) and only after I added the last machine (a shared storage based on FreeBSD) I suddenly noticed that almost every day (usually happens overnight) instead of getting 60Mb/s file transfers, I get 65k/s, it totally actually stalls.

If i check the network speed from the guest with iperf, it is still around 750mb/s, so it is most probably something with disk access. I checked with esxtop, and nothing seems wrong, everything at almost idle.

If I restart the guest in question, it waits at the boot stage and really very very slowly gets up. If I shut the guest down and then power it back up everything works great until it happens again.

I removed one network card (had 4) which was sharing IRQ with RAID controller (are there no more then 16 (0-15) interrupts? I though APIC is enabled in ESXI?) and hoping that it would help, but it did not, I just think that it happens a little more rarely. But this is just a feeling as it can not be triggered by some user action.

I tried to recreate the guest from scratch - of course using the existing data as it is some 800 GB of it - but still the same result. The system disk for this machine is separate, so I created a new one, and installed from iso image, and when configuration is restored the problem happens again. The data is on two different datastores, and all get really slow.

What is very interesting is that another guest (A Mindtouch Wiki virtual machine) gets slow at the same time too. Instead of being able to download an attachment at 30mb/s I am getting 1mb/s max. After the machine in question is restarted, wiki machine starts working again, without touching it at all.

I also upgraded (actually did a clean install) to ESXI 4.0, with hope that it will fix it up, but it does not.

Current list of devices is as follows (irqs):

~ # lspci -p
Bus:Sl.F Vend:Dvid Subv:Subd ISA/irq/Vec P M Module Name
Spawned bus
00:00.00 8086:29a0 1043:81ea V
00:01.00 8086:29a1 0000:0000 11/ 11/0x69 A V
001
00:28.00 8086:283f 0000:0000 11/ 11/0x69 A V
004
00:28.01 8086:2841 0000:0000 15/ 15/0x71 B V
003
00:28.02 8086:2843 0000:0000 10/ 10/0x79 C V
002
00:30.00 8086:244e 0000:0000 V
005
00:31.00 8086:2810 1043:81ec V
00:31.03 8086:283e 1043:81ec 5/ / C V
00:31.05 8086:2825 1043:81ec 5/ 5/0x81 B V ata_piix vmhba1
01:00.00 1000:0060 1028:1f0a 11/ 11/0x69 A V megaraid_sas vmhba2
02:00.00 8086:107d 8086:1082 10/ 10/0x99 A V e1000e vmnic0
03:00.00 8086:10b9 8086:1083 15/ 15/0xa1 A V e1000e vmnic1
05:01.00 8086:107c 8086:1376 3/ 3/0x89 A V e1000 vmnic2
05:02.00 5333:8811 0000:0000 0/ 0/0x91 A V

Most interesting is that on the same host, a server 2008 is running, with Exchange 2007. The speed of this machine is not influenced, it works great all the time. I even made a share on it, to copy a few GB of data back and forth, and it works great.

All the data stores are on a Perc6e raid controller, and if it was the problem, I think that all the guests would be having the same symptoms.

I also replaced the power supply with a new one, but nothing changed.

Is there some way I can try to catch the event that causes the slowdown?

I think that I have some conflict somewhere, but can not find it. When the machine is slow, esxtop is normal. Nothing is running high.

Vmark

Rohail2004 — Fri, 29 May 2009 18:16:00 GMT

anyone using it? how does it work? help would be greatly apperciated

Forecasting Memory Loads - Granted vs Consumed

ehayes — Wed, 27 May 2009 20:00:01 GMT

This is going to be part one of a two part post I figured I put my memory related questions in this thread to solicit feedback on my thoughts. I'll post a second question in regards to cpu later.

I'm using the API to pull back certain data from virtual center for capacity planning purposes. My goal is to be able to establish/forecast when to add additional hardware resources to this environment (add physical memory, upgrade procs, or simply scale servers).

I'm pulling these metric right now to get start and here are my assumptions / questions.

I'm collecting data at a cluster level.

1.) mem.granted.average This will show me aggregate hard limit of all the maximum memory all running virtual machines could use.

2.) mem.consumed.average This will show an aggregate of the memory the running vms are using with the transparent page sharing and internal optimizations.

So my theory is that if I trend off granted I could ensure that I never get into an overcommitted scenario. However using the consumed number shows overall utilization much lower. Would be better to trend off the consumed number to try to push the infrastructure and get more bang for the buck.

Data examples from one of my environments:

8 Host * 32GB Memory = 262144 MB of Physical Memory

mem.granted.average = 269534 MB 102 % ( -7406MB / -2 % ) Overcomitted!!!

mem.consumed.average = 154176 MB / 59%

If I trend off the consumed number I'll be drastically overcommitted. Even if my threshold to add physical resources is 80% I know the easy answer is to act when granted hits 80% but is this the right thing to do...(If I were writing the check)?

I/O performance of vSphere

chethank — Tue, 26 May 2009 18:55:57 GMT

I am a senior performance engineer at VMware. I have spent quite a bit of time on performance analysis of ESX storage stack starting from ESX 3.0. We at VMware are constantly working on improving I/O performance of ESX. I have done few experiments to drive extremely high I/O load on single instance of ESX. With 3.5, I obtained 100,000 I/OPs with an I/O load that is most representative of real applications (100,000 I/Ops) until I ran out of hardware. The performance envelope was pushed furthe with vSphere when we achieved 350,000 I/Ops in an experiment done at EMC labs. I wrote a blog highlighting the results with some details on the experiments in VMware's performance blog VROOM.

Having a well configured I/O system is critical for good application performance. Very often I hear questions from customers on storage performance, choice of virtual disk format - VMFS vs RDM, best practices etc., The answer is simple - follow the best practices that you would normally follow in native world when designing an I/O infrastructure for your application. ESX provides excellent I/O performance and can support even extreme I/O demands from applications as the results discussed in the blogs indicate. VMFS or RDM - you can expect similar performance though RDM can help during certain scenarios which are purely non-performance related.

We can discuss more in this thread. Feel free to post your questions, comments on the blogs or any I/O related issues on this thread. I will try my best to respond. May be some one who has already faced a similar situation will jump in with a solution even we at VMware wouldn't have thought of!

Chethan

CPU Usage (Un-)Optimization (newbie)

TurboThunder — Mon, 25 May 2009 18:48:39 GMT

I am trying out VMWare Workstation to run an old application I have. I want to get rid of my old PC, but the app will not run on Vista64.

Anyway, I have successfully installed Win2K in an VM, and have also successfully run the app. However, performance is poor.

I have an inkling that VMWare is using idle detection to optimize host CPU usage, as I am getting very little CPU usage listed in the host OS's task manager. Sometimes, I can get the CPU usage to be very high and performance of the app is good, but at certain times when the app is checking for input, CPU usage trails off and the app is not very responsive.

As I have 4 cores in the real CPU, I am willing to give up all CPU time to the VM during the entire session for one of the cores.

How can I get VMWare to NOT throttle CPU usage?

I have a feeling the culprit is keyboard polling.

Thanks in advance for any advice.

very slow copy vista(host) to linux(guest)?

bassemfarouk — Thu, 21 May 2009 06:55:36 GMT

Hi ,
i have VMWARE server 2
this is installed on Windows Vista business(3GB RAM)
i have a virtual machine of Oracle Enterprise Linux 5.1 (exactly same as Redhat 5)
when i copy from window vista to linux via WINSCP , it is really very very bad slow copy
any property to tweak ? i really need to copy from/to the 2 OS as fast as possible

Thanks

Packet loss on Sun Fire X4140 / ESXi 3.5u4

HMVM123 — Tue, 19 May 2009 19:30:46 GMT

I have a Sun Fire X4140 system with 4 Nvidia nForce Pro 3600/3050 based NICs. On this system, I have ESXi 3.5 Update 4 running and further, a single VM running a slightly modified CentOS 4.4 distribution.

I've attached one of the interfaces on the VM to a virtual switch connected to one of the physical interfaces. Everybody is in promiscous mode, so that I can install a sniffer application in the VM.

To test the system, I have a separate physical host replay a pcap file to the sniffing interface and I see most of the traffic that's being sent. Unfortunately, there's definite packet loss that grows somewhat linearly with traffic rates. At 1Mbps, I lose around 0.05% of packets. As the rate rises to 200Mbps, the loss approaches 5%. I'm trying to understand the causes for this packet loss.

What's interesting is that VIClient gives me packet numbers that match precisely the number sent. ifconfig inside the VM is where I first see the loss. There are no obvious error messages and the system doesn't appear to be under stress.

Any pointers out there? I certainly notice that ESXi is presenting my VM with an e1000 adaptor interface, while it probably uses forcedeth under the covers to manage the physical interface. Any value in going with vmxnet?

Any help will be much appreciated!

Problems installing Win95 onto VMware fusion =[

stressfulsebby — Tue, 19 May 2009 13:28:22 GMT

hi all,

first of all i am very much a noob at this stuff so please don't decapitate me =P

ok so i obtained an iso image of Win95 and ive attempted heaps of things that i've read on these forums to try and install it.

Whatever i try the VM tells me it can't find an operating system.

Perhaps i havnt properly created a virtual floppy disk or something.

To be honest i don't entirely know what to do to install 95 onto my machine haha.

Could somebody please give me tips on how to install Win95 onto my VM!!? (just incase, i'm running version 2.0.1 on Mac OSX Leopard - black macbook with 2.2GHz intel core duo and 160GB HD space)

any help would be very much appreciated!

Thanks in advance

Performance issues with MS SQL 2005 on a VM

rupeshj — Mon, 18 May 2009 16:02:34 GMT

Good Day All.

I've received complains from the DBA regarding poor performance on a MS SQL 2005 with SP 2

Configuration:

#!/usr/bin/vmware
config.version = "8"
virtualHW.version = "4"
floppy0.present = "false"
nvram = "XXXXXXXXXX.nvram"
deploymentPlatform = "windows"
virtualHW.productCompatibility = "hosted"
tools.upgrade.policy = "manual"
powerType.powerOff = "default"
powerType.powerOn = "default"
powerType.suspend = "default"
powerType.reset = "default"

displayName = "XXXXXXXXXX"
extendedConfigFile = "XXXXXXXXXX.vmxf"

numvcpus = "4"
scsi0.present = "true"
scsi0.sharedBus = "none"
scsi0.virtualDev = "lsilogic"
memsize = "4092"
scsi0:0.present = "true"
scsi0:0.fileName = "XXXXXXXXXX-000003.vmdk"
scsi0:0.deviceType = "scsi-hardDisk"
sched.scsi0:0.shares = "normal"
ide0:0.present = "true"
ide0:0.clientDevice = "true"
ide0:0.fileName = "/usr/lib/vmware/isoimages/windows.iso"
ide0:0.deviceType = "cdrom-raw"
ide0:0.startConnected = "false"
ethernet0.present = "true"
ethernet0.networkName = "Windows Network"
ethernet0.addressType = "vpx"
ethernet0.generatedAddress = "00:50:56:94:1b:c3"
guestOSAltName = "Microsoft Windows Server 2003, Standard Edition (32-bit)"
guestOS = "winnetstandard"
uuid.bios = "50 14 dc e6 63 25 5c 71-30 4b e7 ed 96 c5 46 04"
log.fileName = "vmware.log"
snapshot.action = "keep"
sched.cpu.min = "0"
sched.cpu.units = "mhz"
sched.cpu.shares = "normal"
sched.mem.minsize = "0"
sched.mem.max = "4092"
sched.mem.shares = "normal"
toolScripts.afterPowerOn = "true"
toolScripts.afterResume = "true"
toolScripts.beforeSuspend = "true"
toolScripts.beforePowerOff = "true"

evcCompatibilityMode = "FALSE"
guestCPUID.0 = "0000000a756e65476c65746e49656e69"
guestCPUID.1 = "000006f800010800000022010febbbff"
guestCPUID.80000001 = "00000000000000000000000120000000"
hostCPUID.0 = "0000000a756e65476c65746e49656e69"
hostCPUID.1 = "000006fb000408000004e33dbfebfbff"
hostCPUID.80000001 = "00000000000000000000000120000000"
scsi0:0.redo = ""
tools.remindInstall = "FALSE"
userCPUID.0 = "0000000a756e65476c65746e49656e69"
userCPUID.1 = "000006fb000408000004e33dbfebfbff"
userCPUID.80000001 = "00000000000000000000000120000000"
vmware.tools.requiredversion = "7302"

tools.syncTime = "FALSE"
uuid.location = "56 4d 20 35 ab 87 9c 5f-ca a0 74 cb ec fa 11 fe"
sched.swap.derivedName = "/vmfs/volumes/49b65cd9-a7d109c2-423f-001a645acc3e/XXXXXXXXXX/XXXXXXXXXX-41e0c04f.vswp"

scsi0:1.present = "true"
scsi0:1.fileName = "XXXXXXXXXX-000004.vmdk"
scsi0:1.deviceType = "scsi-hardDisk"

scsi0:1.redo = ""

checkpoint.vmState.readOnly = "FALSE"
checkpoint.vmState = ""

guest.commands.allowSharedSecretLogin = ""
guest.commands.sharedSecretLogin.com.vmware.vcIntegrity = ""

logging = "false"
cpuid.80000001.edx = ""
cpuid.80000001.edx.amd = ""

migrate.hostlog = "./XXXXXXXXXX-41e0c04f.hlog"

cpuid.1.ecx = "------------0-------------------"

VM swap file datastore when using NetApp SMVI- maximum number of VMs?

007Roberto — Fri, 15 May 2009 22:04:53 GMT

When following the best practice with NetApp storage (as stated in TR-3428), configuring swapfiles for VMs on a separate datastore as it pertains to snapshots with SMVI; what is the maximum number of VMs that should have their VMs swapfile on that datastore?

esxtop output not comprehensible

ashutoshmehra — Wed, 13 May 2009 09:23:16 GMT

Hi,

I wanted to study the performance of Transparent Page Sharing (TPS) facility provided by ESX Server.

The setup I have is: ESX Server running 4 guest OS (virtual appliance).

I ran a C program to check the performance of TPS.

1) In first case, the program mmaps a 50 MB file (passed as argument) to its memory and then sleeps for ever.

I started this program on all four guest OS. All four instances of the program map the file which is replicated in each guest OS. In essence, although the four processes are mapping physically different file, but the contents are same. After an hour or so, esxtop showed around 150 MB of memory saving, as expected.

2) In second case, I modified the C program to dynamically allocate 450 MB of data filled with some random values, and then map 50 MB file. So, now the program has 450 MB of random data and 50 MB of file mapped data. Again, I started the program on all four guest OS in same way as in previous run. I expected same amount of memory savings (around 150MB) in, may be, longer period of time, as now the process instance has 450 MB of unidentical data along with 50 MB of similar data (mapped from file). So, TPS may take a bit longer to search and merge the identical pages. But even after leaving it for half a day, it didn't show any kind of expected behavior.

I collected the data for second case and I prepared a chart showing the memory savings over a period of time. I started collecting the data as soon as I started the program on four guests. In th chart, this corresponds to time 3:08. At this stage, memory savings are around 950 MB. An hour later, it increased to 1020 MB (memory savings increased by 70 MB). This value remained constant for some time, and then dropeed sharply to 935 MB. It again increased to 980 MB. This pattern of fall and rise was repeated 3-4 times. All this while, the program that I started was in sleep state.

Is there any way to account for this pattern? Also, why are memory savings in second case just 70 MB, as compared to 150 MB in first case? I expected same amount of memory savings in both cases.

Network Issues large number of Duplicate ACKs

JeffStory — Fri, 01 May 2009 13:51:07 GMT

I am researching network issues on multiple virtual machines and noticing a large number of duplicate ACKs in my network traces. I am seeing this behavior across multiple guest machines across different ESX 3.5 clusters (even being managed by different vCenter servers) in different locations of the network. Some machines are behind a load balancer, some are just trying to communicate with SQL servers on the same VLAN in a local network. My Network Operations guys show that switch utilization (on Cisco 6500s) is less than 10% and show no errors on the switch ports.

Is anybody else seeing a large number of Duplicate ACKs with guest machines on 3.5 ESX servers. To give an example, in a 20 minute network trace (100mb of traffic) I have 5,600 duplicate ACKs. This is with the 2 machines isolated to the ESX server, one being the production server, the other being the machine capturing the traffic.

A little OT: Poor performance on backup

Emmermacher — Mon, 27 Apr 2009 11:51:24 GMT

Hello.

We're running several servers (SLES 10 and W2k) on an ESX 3.5.0 machine. The performance on normal work is well, but during backup the system on a SLES 10 guest is very poor.

We use BackupExec 11.d on a physical W2k server with ralus agent on the linux machines. The configuration for the agents don't anything to increase the performance. The performance on another guest with W2k server is much better.The guests are on a netapp fs2020 filer usins sata disks. The ESX servers are on FSC MX630 machines. The connection use the iscsi protocol. Is there something to increase the performance here?

Thanks in advance for your suggestions.

Best regards.

Dirk Emmermacher

SATA Drives' performance connected to PCI-E x8 vs PCI-X 100

Virtuali3ed — Thu, 23 Apr 2009 13:41:32 GMT

Hi, Since I do not have any prior experience with PCI-X and PCI-E so please advise.

Will there be any pratical performance gain when SATA drives are connected to PCI-E x8 vs PCI-X 100. I understand that PCI-X 100 has a throughput of 800 MBps where are PCI-E x8 has around 3 GBps?

Thanks in advance!!

VM to VM network throughput is really slow.

PingMonster — Wed, 22 Apr 2009 11:42:09 GMT

HI...

I have a throughput problem between 2 VM client machines.

2 of ESX 3.5 (153875)

4 CPU, 24GB Mem, 6 Intel 1GB NICs on each ESX host.

1 vSwitch on 4 teamed NIC

9 VM clients

VM clients are up to date with patches and fixes for windows and vmware tools

Just recently when I copied 3Gb file acrossed between one vm to another vm the network thoughput came down to 5Mb/s

1. Tested with remaining VMs turned off, still same throughput

2. Tested with vm to physical, still same throughput

3. Tested with 'Enhanced vmxnet' adapter both on vm, still same throughput

4.Tested with 'Enhanced vmxnet' or 'Flexible' adapter on both or one each VM, still same throughput

5. Tested with both VM on same ESX host, or different host, still same throughput.

It peaks about 180 to 200Mb/s for few seconds and booooom.....drops to 4~5Mb/s

One other strange things is....it only happens with both Windows 2003, it does not happens between 2003 to 2008 or 2008 to 2008.

Anyone....?

HP ProLiant DL785 G5 achieves #1 virtualization performance result on VMmark benchmark ( http://www.vmware.com/products/vmmark/results.html )

aaklea — Wed, 22 Apr 2009 01:26:52 GMT

HP ProLiant DL785 G5 achieves #1 virtualization performance result on VMmark benchmark

( http://www.vmware.com/products/vmmark/results.html )

NIC Teaming shows two NICs down

ck001 — Tue, 21 Apr 2009 20:33:29 GMT

I'm trying to team a quad port NIC for our production server VLAN and it shows two NIC's up and two with red x's through them. We checked the network side and everything seems OK. Any ideas about this? I'm running ESX 3i v3.5. I'm also more than a little confused about networking. I have a management/vmotion virtual switch, and then I created another vswitch for the production network. However when I tried to add a VMKernel, it would not allow me to set it for the gateway for the production network. Apparently VMKernel can only have one gateway. Should I change the gateway to the prodcution VLAN or keep it on the management network?

pc configuration for Vware

dbasureshkumar — Sat, 11 Apr 2009 10:15:45 GMT

I want to buy the New Pc for Oracle 10g Rac installation(2/3 Nodes or 2/3 or more guest OS ) through VMware which point should consider at hardware level?

If I configure more than 4 GB RAM than Window OS will display only 3 GB RAM?

The old misalignment issue - needed for virtual disks on datastore ?

stucky101 — Thu, 09 Apr 2009 21:04:47 GMT

Guys

I know this has been discussed to death but yet never fully answered. That's why I'm starting yet another, new thread since most others I found are older now and refering to older version of ESX. I will gather here what I think I know and have read in many threads and techpapers and let you guys debate it.

Here is what I know with about 99% certainty.

1. A single un handed to a physical machine running win or lin needs to be aligned to start at block 128 if you decide to partition this lun.

2. If you use LVM2 on the host instead (i.e "pvcreate" the lun, then add to VG, then carve out LV, then put ext3 straight onto LV) you DON'T need to align since no partition was ever created. Not sure why cause doesn't the PV signature that LVM2 writes to the lun also cause an offset ?

3. As of ESX3 using VC when creating a vmfs volume the alignment is taken care of. If you use vmfstools however it's not and you'd have to align manually.

NOW it gets interesting

4. If you hand a real lun straigt to a VM (RDM) I understand that you also need to align according to step 1 or not according to step2 (depends on what you choose to do with that lun). This makes sense to me cause the VM will write the signature onto the lun thus misaligning it if you chose to create a partition.

What I don't understand is that everybody is talking about that as if that was the most common way to hand storage to a VM when in my mind it's not.

Here is my question I still have. Do I need to align anything if I do the following:

1. Create a datastore via VC with one lun in it (ESX alignment taken care of)

2. Create a virtual disk from the datastore and hand it to a VM as a regular disk (no RDM !)

I don't think I have to since I didn't hand an entire lun to the VM - just a virtual slice of it. Isn't that how most ppl assign disks to VM's ? Everybody always talks about luns when really the VM doen'st see luns at all but a datastore instead which could span multiple luns. (except RDM of course)

In this case I don't even know exactly where that virtual disk lives on the lun anyways. I do know however, that it's nothing more than a file on a vmfs volume which is definitely already aligned so why align again inside the VM ?

If someone could answer that for me i'd be so much happier.

Last but not least can someone please define "boot disk" and "data disk" ? I assume "boot disk" means the entire VM ? Or is it refering to "/boot" the boot partition ? By "data disk" I assume they're refering to a virtual disk I assigned to the VM which leads back to the question above as to why I'd have to align it

if its' NOT a full lun (RDM) but just a virtual disk.

I'm hoping this thread can put this issue to bed for good already .

A frustrated newbie...

Cannot see performance graphs from Virtual Center console

Ab2 — Tue, 07 Apr 2009 08:02:41 GMT

I have installed Virtual Center and attached two ESX3 servers to this VC server. Etherything is perfect except one thing - I see nothing on "Performance" tab for these two ESX servers, just "Retrieving chart data..." where graphs are supposed to be.

But if I connect (with VI client) directly to ESX servers - I can see performance data.

Please, give a hint. Why this happens and what should I do to get the ability to see performance data via VC server?

how memory.swapped is calculated?

arpitmpatel13 — Tue, 07 Apr 2009 06:43:07 GMT

Hi,

I am using ESX3i. I am monitoring memory performance counters.

In memory performance counters there are 4 counters which are related to swap usage of memory (for VM), which are as follows.

in that, the memory.swapped is always the difference of memory.swapout and memory.swapin

i.e. memory.swapped = memory.swapout - memory.swapin

So, here I am confused @ the meaning of memory.swapped as it says "memory in kb that is swapped".

Could anyone please give some details about how these memory.swapped and memory.swapin and memory.swapout are related?

Thanks

Arpit

Datastore Working Space

atarso — Mon, 06 Apr 2009 18:50:06 GMT

Is there a recommended or required amount of "working space" (i.e. elbow room) for datastores? Just wondering if there would be a performance hit unbeknownst to us in the event the datastore was down to say...1GB out of 800GB for example. I searched Google and this site but I can't find something towards this aside from the users guide but it doesn't give a definition toward this either (just indicates how to create them, etc.).

Suspect my logic may be flawed there though! Any insight would be appreciated :o)

Problems logging on when reverting to snapshot

shlammed — Fri, 03 Apr 2009 17:42:00 GMT

I have deployed several Windows 2003 VM's for developers here and many have no issues. However, once in a while a few of them have this issue:

When they revert to a snapshot, they cannot logon to the network. The fix is to logon as local administrator and re-add the VM to the domain.

Has anyone had this issue before? I have no idea what is causing it. It's inconvenient for myself and them as they need my credentials to re-add to the domain.

Slow read speeds from VM network shares

microkid — Wed, 01 Apr 2009 18:56:25 GMT

Posted this first in ESXi, but this is the better forum.

Hi. I built a whitelist ESXi server, consisting of:

Intel DG33TLM mobo
Intel E8400 CPU
4GB RAM
Samsung 320GB SATA disk (via ICH9 onboard) (for storing VM's)
Dell Perc 5/i RAID controller with 256MB cache and battery
3x1TB in RAID-5 (data volume, 2TB netto)
Intel Pro/1000 PT dual gbit nic

I have a VM that acts as file server. The VM is on the Sata disk, and there is a 1TB vmdk on the Raid volume which contains all my data.

All installed fine, VM's are running fine too. But I seem to have some unexplained speed problems over the network. When reading from a share on the VM the speeds are very low: in order of 20-40MB/sec. Write speeds are a bit better, approx 50MB/sec. When I run the atto disk bench in the VM, the speeds are normal : approx 100MB/sec read and 200MB/sec write (thanks to the raid controller). Also wehn I copy large files between VM's speeds never get higher than 30MB/sec. Much too slow on any Sata or Perc controller.

So the speeds are only slow when the data is copied withing the VMs. The funny thing is: before installing ESXi, this pc ran W2K8 and disk speed over the network were great, approx 80MB/sec. Only after installing ESXi and running a virtual W2K8 the speeds dropped dramatically.

I tried:

setting the NIC to 1000Mb full. No change.
installing the latest VMtools. No change.
Installing the U4 update. No change
played around with tcp/ip offload in the VM. No change
Using another Gbit nic (Realtek). No change.
Using W2K3, W2K8, XP, Vista. No change

Any clue why the network speeds are so slow outside the ESXi box?

Different memory overhead statistics in VIC

david_drew — Wed, 01 Apr 2009 17:55:44 GMT

There are two different statistics labeled "Memory Overhead", and I'm trying to figure out what the difference is.

There's a Memory Overhead field in the VIC Summary screen, in a sample VM the value is: 112.65MB

But if I look at the performance counters for Memroy Overhead (real time, daily, monthly--it doesn' t matter), I see values around: 58.82 (when converted from KB to MB).

Does anybody know the difference between the fields, and ideally, how to convert one to the other?

Thanks,

-David

capture the LUN utilization

sony1001 — Thu, 26 Mar 2009 02:45:16 GMT

Hi All,

In my production environment there are lot of of ESX servers are running with shared LUN's. My question is How to capture the LUN Utilization or is there any scipt available to capture the Lun utilization ? if any body would like to share the script that would be great. Requesting all you provide your valuable comments, Suggestions...

Thanks in Advance!!.....

ESX3.5 W2008 client VM performance is slow compared to W2003

DSMXP1 — Wed, 25 Mar 2009 14:56:44 GMT

We are running ESX with Sans, when I compare my VM's the ones taht are runing w2003 compared to w2008 , the 2008 machines seem be sucking up resouces and run slower, the processor hit's 100% at times running minimal applications.

I am running and appache web page off the 2008 that has a search filed to search files from a directory, when I run the search from the appache on the 2008 machine it appears slow.

If I run the same appache page from a test Work station, pointing to the same directory its almost 3 times as fast. both machines are w2008 servers.

Why would the VM on the ESX be so much slower. Any ideas I could hunt for.

VCB performance problem relays on Windows 2003 proxy disk with HP P400 raid controller?

icarium — Tue, 17 Mar 2009 13:31:57 GMT

Hi all,

using Windows 2003 R2 sp2, all updates, on a HP DL 380g5 with a 5420 QuadCore, 4 GB ram. System runs on a 72 GB SAS, and 6 x 72 GB SAS 10k RPM disk as raid 0 for backup VM per FibreChannel. Attached at a HP P400 with FW 5.22 and 6.14.0.32b drivers.

The problem is the performance of the raid system. When using 1 disk, you get ~70 MB/s, using 6 disk in raid 0, its about 100 MB/s. All chaches are on, 50%read/50% write, disk caches are turned on etc on raid controller. The next thing was the idea, to change the stripe size of the raid, standard was 64 KB. Going up and down does not change anything big. The second option is changing block /cluster size when formatting it with Windows from standard 4 KB to 64 KB, in conjunction with setting stripe size to 16 KB. So when accesing the raid, with 64 KB size, it takes 4 disk accesing and performance increased to app. 150 - 200 MB/s maximun. But using 6 disk the performance should be higher. So setting stripe to 8 KB and 64 KB when formatting, this does not change.

How to improve the performance anymore?

When using benchmark programs, IOmeter or HD tune, and setting the block size to 1 MB, you get 300 MB/s, using 8 MB its about ~400 MB, what could be the maximum. But for VCB, I didn't find anything to change the block size this application is using.

Any ideads appreciated.

Thanks

Icarium

VDM Agent vs RDP for dialup user?

danielsdavidr — Mon, 16 Mar 2009 14:46:15 GMT

I have a user on a 56K dialup line, connecting to office via PPTP VPN. User is needing to access an applicaiton on a PC/VM however I was wondering if it would be better to have them setup to RDP to the PC/VM or create a VDM connection and have them use VDM Agent. I was just wanted to see others inputs on this to see if they have had complications either way.

ESX 3.5 u3 very slow disk write performance

Merlothun — Sun, 15 Mar 2009 13:10:09 GMT

Hi everyone!

Here's our problem:
We hava a brand new server with esx3.5 u3 with the following components:
-MB: Intel S3210SHLC
-CPU: Xeon 3110
-mem: 6 GB DDR800 ECC
-RAID controller: LSISAS3041E-R
-HDD: 2x500 GB Seagate SATA2 16 cache (local RAID1)

We tested the performances:

virtual win XP
read write
network file copy with a physical host (booth Gb) 50 MB/s 6MB/s

server bare metal speed
read write
knoppx live CD, command dd 130 MB/s 130 MB/s
win2003 server (not virtual) network file copy 75 MB/s 65 MB/s

knoppx live CD, iperf with physical host (through Gb) 750 Mbps

So the problem is, that the one XP VM under ESX have very poor disk write speed. Compared with the bare metal performaces I can't understand, what can be the problem.
I read a lot of KBs, forums without luck.

Did anybody met something like this?

I appreciate any ideas.

Thanks in advance!

ps: we tried XenServer 5 too, we experienced the same performance problem..