Virtual processors

Hi,
Was wondering if someone would be able to shed some light on a few questions I have regarding virtual processors on a T5220 server?
1.) I have a perl program (not threaded) that runs a certain amount of time on a Intel based HP server with a 2.2 Mhz processor. The same script runs about 4 times slower on a T5220, 8 Core, 64GB RAM server.
Now I see that there are 64 1.2 Mhz virtual processors on this server and since the perl script is not multi-threaded, I assume it uses just one of the virtual processors? And running at 1.2 Mhz, it is much slower than my old Intel box, is that right?
2.) Now the question is, can these virtual processors be reconfigured to less, but more powerful processors or am I missing the boat here?
3.) If I am missing the boat, are there any suggestions on how I can speed up my perl script on this powerful machine?
Regards,
Boela.
Edited by: Boela on Aug 27, 2009 4:46 AM

Hi Kosie
As you have seen, the coolthreads servers are very good at running multi-threaded processes or many single threaded processes, but poor when it comes to running one single threaded processes quickly. These machines work best on large multithreaded workloads, such as high volume webserving.
In your case it sounds like your server has 8 cores and each core can run 8 concurrent threads, this gives you a total of 64VCPUs.
In answer to your questions.
1) If your perl program is single threaded then yes it will only use one of the 64 Virtual CPUS. If you look at prstat or similar, the most CPU that process will use on the server is 100/64 = approx 1.6%. To use the rest of the processing power available on the T5220 your script will need to become multithreaded. And depending on what it does, that may not be easy. So you may have to just accept that there is no easy way to make it run faster on the T5220 as it stands. When it comes to single threaded performance the T5220s are typically slower than a modern x86 machine.
2) No you cant configure the VCPUS to be fewer in number and greater in processor speed.
3) To make your script run faster you will need to either tune it or if possible make it multithreaded, or split the workload up into chunks and run multiple instances of the perl script.
Sorry, but no easy solution to your problem.

Similar Messages

  • How much Number of virtual processors for VMs

    Hi,
    there is lots more information about logical processors, cores and CPU and their definitions, on the internet. also here in this forum.
    In order to make sure i am not lost on internet, it is better to take experts advice like you.
    I have simple question read on:
    I have a Hyper-v cluster host with 4 sockets (4 CPU's) with 6 cores each. so total are 24 cores.
    I can see in hyper-v host Logical processors are 48. (which may be 1 proc = 12 logical processors, so 4 proc x 12 = 48 logical processors)
    now how much virtual processors I can assign to each VM. I know I can assign more logical processor depending on application requirement. and it affects application performance. For each VM, it is showing maximum i can select 48 virtual
    processors.
    is it true I can select any number of virtual processors in each VM? is there any limitation?
    is it that if i have 8 VMs if i set virtual processors to 6 then it will occupy all logical processors (48 in my case) and i  will not be able to allocate logical processor to new VMs?
    is it that if i have 10 VMs and all set to 10 virtual processors? 
    thank you

    Quite honestly - this math really does not matter that much with modern processors. (a combination of speed and core count, and advances in CPU scheduling in the hypervisor)
    You have to have a VM with many vCPU to get into any point of contention. 
    In practice, It is very difficult to over-subscribe your CPU resources without hitting some other bottleneck first.
    Set the vCPU count where it is appropriate for your workload.  That means testing your workload to see where adding additional vCPU to a VM provides no benefit in throughput or speed or capacity for your application.
    A select few applications are written in a way that there is a linear correlation between CPU count and throughput.
    There is an old point about not assigning so many vCPU to a VM that the VM could end up with more than one execution thread on the same processor (core).
    You have 24 cores.  and each has 12 logical processors (execution threads).  So, with zero CPU over-subscription you could have 288 vCPU assigned to any number of VMs.  To expand my point in the previous paragraph - you can easily / safely go
    over this count of 288 - what I am referring to above is an old guideline that would guide you to not assigning more than 288 vCPU to a single VM - ensuring that no single execution thread is trying to service more than one vCPU of the same VM.
    Brian Ehlert
    http://ITProctology.blogspot.com
    Learn. Apply. Repeat.

  • Physical vs virtual processors

    Hi,
    What are virtual processors. How many virtual processors exist per physical processor.
    Thanks.

    Available documentation, under:
    (Virtual core/physical socket)
    http://www.oracle.com/us/corporate/pricing/cloud-licensing-070579.pdf
    (Virtual threads)
    http://labs.oracle.com/scalable/pubs/WIOSCA2006.pdf
    Bye.

  • Determine processor ids for virtual processors on the same phys. processor

    Hi!
    Is there an API to determine the processor ids of all virtual processors that belong to a physical processor (for example the cpu ids for all cores that are on one socket in a multi-core architecture)?
    I'm looking for a solution especially for UltraSPARC and SPARC64 CPUs.
    There is the command "psrinfo" which, when called with the arguments "-p -v" procuces an output like:
    The UltraSPARC-IV+ physical processor has 2 virtual processors (0, 16)
    The UltraSPARC-IV+ physical processor has 2 virtual processors (1, 17)
    The UltraSPARC-IV+ physical processor has 2 virtual processors (2, 18)
    The UltraSPARC-IV+ physical processor has 2 virtual processors (3, 19)
    On this systems, processor ids 0 and 16 belong to the same socket, as well as 1 and 17 and so on. On larger systems (e.g. a SunFire6900), processor ids 0 and 512 are the ids of the two cores on the same socket, 1 and 513 and so on.
    Is there any API that allows to determine this information, which virtual processor ids are ids for processors on the same physical processors?
    Thanks,
    Nick.

    On recent versions of the OS, that information is exposed through 'kstat'. Look at the 'cpu_info' stuff. Each VP should have a 'core_id' as well. If two CPUs have the same core_id, then they're multi-threading cores exposed as multiple processors.
    Then I think (but not as sure) that 'chip_id' is similar, but refers to packaging assembly and is similar to the information you're seeing on psrinfo.
    Darren

  • "Load Average" vs "virtual processor"

    Hi,
    I have one question regarding the understanding of “load average” in a platform with virtual processors.
    Suppose in this situation:
    Total number of physical processors: 1
    Number of virtual processors: 32
    Total number of cores: 4
    Number of cores per physical processor: 4 Number of hardware threads (strands or vCPUs) per core: 8
    In the load average, does 32 indicates a full load? Or it is 4 represents a full load?
    Many thanks,
    MDING

    The kernel scheduled according to how many CPUs it "thinks" it has. Which would be 32 in this case I believe. So a load average of 32 would indicate a full load if you are looking at it within the virtual environment. It would be different if you are looking at the host platform...

  • Hyper-V Hypervisor Virtual Processor\\% Guest Run Time occasionally has bad values

    I'm accessing PerfMon data via RegQueryValueExW () calls in a C++ program to collect a variety of metrics. I've noticed that
    Hyper-V Hypervisor Virtual Processor\\% Guest Run Time occasionally contains incorrect values, as shown both in my collected code as well as in PerfMon. (Also % Total Run Time, and possibly all the metrics.)
    My test environment is running 2008 r2 standard, service pack 1, on an Intel Zeon CPU (if it's relevant), and is hosting four guests, none very active (_total = .449, the rest = .030, .513, .185, .194, .341).
    The problem is that occasionally the first, least active, guest will spike well above 100% (as an example, 6300%).
    When walking through my code, the issue seems to coincide with the order of the subscripts being returned from RegQueryValueExW () changing, where the first, least active subscript moving to the second location, and getting an unexpected value. When it returns
    to the first spot in the series on a subsequent call it once again has an expected value. Oddly enough the new "first array element" has the expected value even after the shift.
    Actual returned values for a 30 second collection cycle, values are shifted by 1000:
    GuestRunTime[hyper-vtest_0]; current = 1211813.198, previously = 1211804.192
    GuestRunTime[win2k3EntR2_-_prometheus_3]; current = 18401087.260, previously = 18400964.194
    (interval)
    GuestRunTime[hyper-vtest_0]; current = 1211822.625, previously = 1211813.198
    GuestRunTime[win2k3EntR2_-_prometheus_3]; current = 18401201.918, previously = 18401087.260
    (interval, and then the array order changes)
    GuestRunTime[win2k3EntR2_-_prometheus_3]; current = 18401327.210, previously = 18401201.918
    GuestRunTime[hyper-vtest_0]; current = 601735.139, previously = 1211822.625
    (interval, and then the array order changes back)
    GuestRunTime[hyper-vtest_0]; current = 1211841.390, previously = 601735.139
    GuestRunTime[win2k3EntR2_-_prometheus_3]; current = 18401472.446, previously = 18401327.210
    (interval)
    GuestRunTime[hyper-vtest_0]; current = 1211850.605, previously = 1211841.390
    GuestRunTime[win2k3EntR2_-_prometheus_3]; current = 18401577.243, previously = 18401472.446
    As you can see, the value supplied when the array shifts is rather wonky, but returns to an expected value after the array shifts back: 1211813.198 -> 1211822.625 -> 601735.139 (?) -> 1211841.390 -> 1211850.605.

    Thanks Deepak for your response.
    1) Firewall is turned off
    2) Below is the output from R3trans -d command:
    C:\Windows\system32>r3trans -d
    This is r3trans version 6.24 (release 741 - 04.02.14 - 20:14:02 ).
    unicode enabled version
    2EETW169 no connect possible: "DBMS = SYBASE --- "
    r3trans finished (0012).
    3) The latest dbsl patch that I see on SMP is indeed 35 (Our Kernel is 7.41 64-bit Unicode)

  • SAPB1 in virtual machine?

    Hello!
    Has anyone had a success story running SAP B1 in a virtual machine, particularly virtualbox? Here is our problem: sql or SAP installer is very inefficient and stalls the installation process at 100% cpu for hours.
    Our setup. Host server:
    2 x Intel(r) Xeon(r) CPU E5620  @ 2.40GHz (2 x 4 cores, 2 x 8 virtual processors)
    24GB RAM
    6 x 1TB SATA-3GB 7200 rpm in RAID5 (raidz1)
    Software: Opensolaris, SunOS 5.11 snv_134b
    Virtualbox 3.1.8
    SAPB1 Virtual Machine (more details - at the bottom):
    Hardware (virtual)
    4 cores, 8GB RAM, 32MB video ram, 2 HDDs:
    C:\ - host: VMDK 80GB, dynamically allocate storage, on RAID5/raidz1 volume, guest: NTFS
    E:\ - host: VDI 200GB, dynamically allocate storage, on the same raidz1, guest: NTFS
    Software:
    Windows Server 2008 R2 64bit (180 days evaluation license)
    SQL Server 2008 R2 (180 days evaluation license)
    SAP B1 8.81
    Situation:
    SQL server is configured by SAP installer. When installing SAP Demo database the VMu2019s cpu usage stays at 100% with sqlserver process at 18%, RAM usage is at only 40% (60% free), the installation process takes 6-7 hours to complete. SQL data and log files are placed in the same C:\ drive as the guest operating system. Disk activity shown by the host system is on average 300KB/s writes. The load on the host machine is 10% cpu for the copy of VirtualBox running this VM but no other cpu- ram- or disk-intensive tasks running at all. Earlier tests showed 30MB/s sustained writes on raidz1 but significantly faster in bursts due to effective ZFS caching in hostu2019s RAM. For what I can tell, there is plenty of disk bandwidth, host cpu and ram and guest ram.
    Thereu2019s also a weirdest thing: CPU figures on guest donu2019t add up: performance monitor shows 100% cpu used on all four cores but when you look at individual processes the biggest one is sqlserver at 18%, SAP setup is 0-3%, taskmanager itself is 6-12%, the rest is negligible including IDLE process at 0-2%. The machine is clearly very busy with slow screen updates etc. but the sum of all running processes plus idle process is at 40-50%. Once install of the demo database  sucessfully finished many hours later cpu usage came back to normal - 95-97% idle.
    Suspicions:
    1. SAP installer and/or production executables are compiled so that they arenu2019t happy at all under a hypervisor or only under virtualbox in particular. Earlier trial installation of the an old, 2004, version of SAPB1 on the same VM didnu2019t show this problem although we didnu2019t specifically test performance, it was just an early test installation. Slow u201Cby designu201D?
    1a. SAP is not happy with evaluation versions of Windows Server or SQL. Slow u201Cby designu201D?
    2. Placing demo DB files and logs on a different virtual drive (E:\) or an iSCSI drive will dramatically improve performance of SQL - This is next thing we are going to try.
    2a. VirtualBoxu2019es realisation of VMDK (VMware disk format) is affecting performance.
    2b. The format of virtual disks was chosen to dynamically allocate storage which is, although it seems plenty of spare host disk performance is available is still inefficient. But then, why CPU usage is the problem? Is that caused by badly written .exe files that do not return spare cpu cycles while waiting for disk operations?
    Many thanks for any tips!
    More on VM:
    Guest OS:        Windows 2008 (64 bit)
    Memory size:     8192MB
    VRAM size:       32MB
    Number of CPUs:  4
    Synthetic Cpu:   off
    CPUID overrides: None
    Boot menu mode:  message and menu
    Boot Device (1): Floppy
    Boot Device (2): DVD
    Boot Device (3): HardDisk
    ACPI:            on
    IOAPIC:          on
    PAE:             on
    Time offset:     0 ms
    Hardw. virt.ext: on
    Hardw. virt.ext exclusive: on
    Nested Paging:   on
    VT-x VPID:       on
    Monitor count:   1
    3D Acceleration: off
    Storage Controller Name (0):            SATA Controller
    Storage Controller Type (0):            IntelAhci
    Storage Controller Instance Number (0): 0
    Storage Controller Max Port Count (0):  30
    Storage Controller Port Count (0):      2
    Storage Controller Name (1):            IDE Controller
    Storage Controller Type (1):            PIIX4
    Storage Controller Instance Number (1): 0
    Storage Controller Max Port Count (1):  2
    Storage Controller Port Count (1):      2
    SATA Controller (0, 0): /export/home/admin/.VirtualBox/HardDisks/RAID/disk1.vmdk
    SATA Controller (1, 0): /export/home/admin/.VirtualBox/HardDisks/RAID/disk2.vdi
    IDE Controller (0, 0): /opt/VirtualBox/amd64/additions/VBoxGuestAdditions.iso

    While the other suggestions here have merit, I wonder what can be done with your current solution...
    Why are you running such an old version of VirtualBox?  Version 4.1.4 is available for Solaris machines; I would think you could use that under OpenSolaris (although I am unfamiliar with the details of OpenSolaris, as I haven't run it).
    The dynamically-allocated disk is the second most likely culprit in the performance issues, right behind the older version of VirtualBox...  I have used the dynamic allocation many times for file servers and application servers, but a database server needs to have the highest disk I/O performance possible.  I don't like to see my client application waiting on the database to allocate new disk space which is waiting on the hypervisor to provide additional disk...  I think you see my point.
    You might take a snapshot or clone the VM and test on another machine, even to the point of cloning the existing virtual hard drive to one which was created as a statically allocated disk.
    Those would be the first two items I would try out.  Let us know how successful you are; I'd love to hear what seems to be the bottleneck, as we are considering placing our B1 services inside a VM.  We are considering VirtualBox, VMware, and Citrix Xen at the moment.
    Rob

  • Unable to create filesystem (mkfs.ext4) on large 2TB GPT virtual disk using Linux VM.

    I am unable to create a file system on a large (> 2TB disk) virtual disk for a Linux VM.  I can create the disk, attach it to the VM, partition it with "parted", but I cannot run mkfs.ext4.  Details below.
    Hyper-V 2012 Core (w/ all Windows/Microsoft updates as of 4/19).
    CentOS 6.4 VM w/ 4 virtual processors, 4GB RAM, and 3 dynamic drives: 
    /dev/sda  100GB IDE dynamic vhdx
    /dev/sdb  75GB IDE dynamic vhdx
    /dev/sdc  10TB SCSI dynamic vhdx
    Using parted, created 500GB partition on the 10TB drive (/dev/sdc1). 
    (parted) select /dev/sdc
    Using /dev/sdc
    (parted) print
    Model: Msft Virtual Disk (scsi)
    Disk /dev/sdc: 11.0TB
    Sector size (logical/physical): 512B/4096B
    Partition Table: gpt
    Number  Start   End    Size   File system  Name                Flags
     1      1049kB  500GB  500GB               production_archive
    then run: mkfs.ext4 /dev/sdc1
    repeating error on console from mkfs.ext4:
    INFO: task mkfs.ext4:2581 blocked for more than 120 seconds
    "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
    Runaway error in var/log/messages until my /var system filled up - 25G worth of it:
    -rw-------. 1 root root 25085329408 Apr 19 23:15 messages
    Apr 19 17:39:28 nfs2 kernel: sd 4:0:0:0: [sdc] Sense Key : No Sense [current]
    Apr 19 17:39:28 nfs2 kernel: sd 4:0:0:0: [sdc] Add. Sense: No additional sense information
    Apr 19 17:39:28 nfs2 kernel: hv_storvsc vmbus_0_13: cmd 0x93 scsi status 0x2 srb status 0x6
    Same problem happens when running "mkfs.ext4 -E lazy_itable_init=1 /dev/sdc1"

    Hi,
    Thank you for your post.
    I am trying to involve someone familiar with this topic to further look at this issue.
    Lawrence
    TechNet Community Support

  • Solaris 10 not 'seeing' all processors on e450

    When I do a psrinfo I only see 0 1 3
    During boot the system says 4 x 296 mhz ultrasparc...
    I only see 3 processors when looking into the CDE sun console app. as well.
    Is this normal? I did a boot -r but nothing changed.
    Thanks

    SUNW,Ultra-4
    SUNW,builtin-drivers (driver not attached)
    SUNW,pll (driver not attached)
    SUNW,envctrl, instance #0
    dev_path=/pci@1f,4000/ebus@1/SUNW,envctrl@14,600000:envctrl0
    SUNW,UltraSPARC-II (driver not attached)
    SUNW,UltraSPARC-II (driver not attached)
    SUNW,UltraSPARC-II (driver not attached)
    SUNW,ifp, instance #0
    dev_path=/pci@6,4000/SUNW,ifp@2:devctl
    dev_path=/pci@6,4000/SUNW,ifp@2:fc
    SUNW,ifp, instance #1
    dev_path=/pci@6,4000/SUNW,ifp@3:devctl
    dev_path=/pci@6,4000/SUNW,ifp@3:fc
    Here is the psrinfo -v
    Status of virtual processor 0 as of: 07/11/2005 10:45:27
    on-line since 07/11/2005 10:02:27.
    The sparcv9 processor operates at 296 MHz,
    and has a sparcv9 floating point processor.
    Status of virtual processor 1 as of: 07/11/2005 10:45:27
    on-line since 07/11/2005 10:02:28.
    The sparcv9 processor operates at 296 MHz,
    and has a sparcv9 floating point processor.
    Status of virtual processor 3 as of: 07/11/2005 10:45:27
    on-line since 07/11/2005 10:02:28.
    The sparcv9 processor operates at 296 MHz,
    and has a sparcv9 floating point processor.
    I ran the diagnostic boot and nothing strange came up.
    hmmm...

  • Physical processor to Logical Processor ratio

    Hi Experts !
    I came across this old blog post on Physical processor to Logical processor calculation.
    http://blogs.msdn.com/b/virtual_pc_guy/archive/2010/08/13/using-powershell-to-find-the-virtual-processor-to-logical-processor-ratio-of-hyper-v.aspx
    The straight formula I see here is
    Write-Host (@(Get-WmiObject -Namespace "root\virtualization\v2" -Class MSVM_Processor).count / (@(Get-WmiObject -Class Win32_Processor) | measure -Property NumberOfLogicalProcessors -Sum).Sum) "Virtual Processor (s) per logical processor"
    -ForegroundColor yellow
    As per this formula, My Hp Blade server BL460c with 2 X Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz, 2600 Mhz, 8 Core(s), 16 Logical Processor(s) can have 2.125 Virtual Processor (s) per logical processor
    I have 32 Logical Processors and can have have 68 Virtual Processors in this case.
    Is this assumption correct?
    Is this formula applicable for even an AMD processor?
    Cheers.
    Optimism is the faith that leads to achievement. Nothing can be done without hope and confidence.
    InsideVirtualization.com

    Thanks Tim.
    I got your point.  Its purely on the load on individual VM.
    Now we are in the process of implementing Exchange 2013 and as per the technet article, it clearly says  Virtual to Logical processor ratio recommended for Exchange is 2:1. This made me to think again on the processor ration now.
    Many hardware virtualization products allow you to specify the number of virtual processors that should be allocated to each guest virtual machine. The virtual processors located in the guest virtual machine share a fixed number of logical processors
    in the physical system. Exchange supports a virtual processor-to-logical processor ratio no greater than 2:1, although we recommend a ratio of 1:1. For example, a dual processor system using quad core processors contains a total of 8 logical processors in
    the host system. On a system with this configuration, don't allocate more than a total of 16 virtual processors to all guest virtual machines combined.
    Its a recommendation from the individual product group and only applicable here.
    Thanks for the help.
    Optimism is the faith that leads to achievement. Nothing can be done without hope and confidence.
    InsideVirtualization.com

  • Processors - how many per VM?

    Hi,
    I have an HP G7 with 2 CPU's. Each CPU has 4 cores (8 in total).  What is best practice for allocating CPU to 2 VM's running on one server?  Currently I have 4 virtual processors allocated to each VM.  I'm looking to get the best performance
    out of both.
    Any advice?  Thanks, Matt.

    Hi,
    Refer below link.
    http://blogs.technet.com/b/virtualization/archive/2011/04/25/hyper-v-vm-density-vp-lp-ratio-cores-and-threads.aspx
    Regards,
    MD Disclaimer: The opinion expressed herein are my own knowledge. Deploy this at your own risk. Whenever you see a helpful reply, just click on “Propose As Answer” / “Marked As Answer” and please do "VOTE".

  • Virtual Vista SP or Server 2008 SP1 x86 problem

    New issue:
    When running Lookout 6.1 server process on a Vista SP1 (x86) or Win2k8 (x86) as a virtual machine under Win2k8 (x64) Hyper-V, if you set multiple virtual processors on the virtual machine, Lookout freezes after short random time (minutes to hours).
    Symptoms are frozen screen and 0% CPU on Lookout.exe.  Lookout.exe can be killed only by shutdown or restart and not through task manager "End Process", or stop-process command.
    If virtual machines are set to single virtual processor, all is OK.
    Please look into fixing this....here's why:
    My current project with lots of client hypertrends taxes the single processor machine by causing nicitdl5.exe to use nearly 100% CPU when clients have high data demand such as when changing Hypertrend screens, etc.  With multiple client machines demanding data from Citadel on the same server at the same time, things can get ugly.
    For the short time of testing between freezes on the multi-processor virtual machine servers, this problem goes away and the server is much faster at getting data out to clients and the CPU utilization is MUCH less, even overall.  Very smooth serving on multi-processor VM.....
    Except of course...you can't keep the damn thing running....j
    Thanks,
    Ed

    Update:
    A response on Microsoft 2k8 Virtualization forums had a good point as follows.  LabVIEW/DSC 8.5 and Lookout were installed on these VM's when configured for a single processor.  It was after they were installed and working that I switched to a second virtual processor and caused the issue.  So maybe a HAL list change is what Lookout can't handle?
    So I will next test with a clean VM running 2 virtual processors prior to installing LavVIEW/DSC/Lookout and see if problem goes away.  I'll get back here with results within a day or two.
    Ed

  • Creating a Linux virtual machine on dynamic vhdx file creates huge file on formating

    Hello,
    we just created a virtual machine on Windows Server 2012 R2 on a dynamic vhdx file.
    When the operating system (Redhat 5) formats the filesystem (100GB of size), then the vhdx file expands to 21 GB, although no data is written to the disk.
    Did someone experience a similar issue? And how to avoid such space waste. If we write additional data in the virtual machine, then the vhdx expands again and gets even bigger then the 21 GB.
    Thanks,
    Thomas

    ~~
    The RedHat 6.5 is not supported guest system,
    ~
    RHEL 6.5 in VM on Hyper-V is even _certified_ solution
    ~
    P.S. About .vhdx itself  and BlockSizeBytes 1MB  ( 32Mb as default )  see next message
    P.P.S.
    ~ ~
    "So RedHat itself supports the virtual machine, but Microsoft not?"
    Depends upon how you define the word 'support'.  Microsoft has never supported any Linux operating system in the manner that it supports software it has written.  In this case, support means providing fixes to the software code.  That is pretty
    standard across the board.  RedHat does not support Microsoft applications or software, either.
    However, if you mean the Microsoft supports the running of the RedHat operating system within Hyper-V, the answer is yes.  In those cases where it is not obvious that a problem is occurring in Microsoft's integration components, it will work with RedHat
    to determine where the error lies.  If the cooperative effort between Microsoft and RedHat determines the problem is with Microsoft code, Microsoft will fix (support) it.  If it is determined that the problem is not with Microsoft code, RedHat will
    make the determination as to whether or not to fix (support) it.
    .:|:.:|:. tim
    ~ ~
    https://www.redhat.com/promo/svvp/
    ==
    Red Hat Enterprise Linux and Microsoft Windows Virtualization Interoperability
    In February 2009 Red Hat and Microsoft signed reciprocal agreements
    to enable cross-certification of the companies' virtualization
    platforms.  . . .   . As a result, customers are able to
    deploy heterogeneous virtualization environments comprising Microsoft
    Windows and Red Hat Enterprise Linux with full technical support
    available from both companies.
    The completed certifications include:
    Certification of Windows Server 2008 Hyper-V, Microsoft Hyper-V
    Server 2008, Windows Server 2008 R2 Hyper-V and Microsoft Hyper-V Server
    2008 R2 host with Red Hat Enterprise Linux 5.2, 5.3, and 5.4 guests.
    Customers with valid support agreements receive cooperative technical support for:
     . . . , and
    running Red Hat Enterprise Linux virtualized on Windows Server 2008 Hyper-V.
    ==
    Hyper-V Supported Virtual Machines and Guest Operating Systems
    ==
    Supported guest operating systems
    Server guest operating system (Linux) on Hyper-V in Windows Server 2008 R2
    Editions
    Virtual processors
    CentOS 5.5-5.9
    x86 edition and x64 edition
    1, 2, 3, or 4
    CentOS 6.0-6.4
    x86 edition and x64 edition
    1, 2, 3, or 4
    Red Hat Enterprise Linux 5.5-5.8
    x86 edition and x64 edition
    1, 2, 3, or 4
    Red Hat Enterprise Linux 5.9
    Red Hat certified.
    x86 edition and x64 edition
    1, 2, 3, or 4
    Red Hat Enterprise Linux 6.0-6.3.
    x86 edition and x64 edition
    1, 2, 3, or 4
    Red Hat Enterprise Linux 6.4
    Red Hat certified.
    x86 edition and x64 edition
    1, 2, 3, or 4
    SUSE Linux Enterprise Server 10 with Service Pack 4
    x86 edition and x64 edition
    1, 2, 3, or 4
    SUSE Linux Enterprise Server 11 with Service Pack 1, Service Pack 2 or Service Pack 3
    x86 edition and x64 edition
    1, 2, 3, or 4
    Ubuntu 12.04 and 12.10
    x86 edition and x64 edition
    1, 2, 3, or 4
    Ubuntu 13.04 and 13.10
    x86 edition and x64 edition
    1, 2, 3, or 4
    Oracle Linux 6.4
    Note
    Oracle Linux is only supported when running the Red Hat Compatible Kernel.
    x86 edition and x64 edition
    1, 2, 3, or 4
    ==
     +
    ==
    CentOS 5.7 – 5.8
    Drivers: IDE, SCSI, and networking
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are distributed through a download available at
    http://www.microsoft.com/en-us/download/details.aspx?id=34603
    CentOS 5.9
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    CentOS 6.0 - 6.3
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are distributed through a download available at
    http://www.microsoft.com/en-us/download/details.aspx?id=34603
    CentOS 6.4
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    Red Hat Enterprise Linux 5.7- 5.8
    Drivers: IDE, SCSI, and networking
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are distributed through a download available at
    http://www.microsoft.com/en-us/download/details.aspx?id=34603
    Red Hat Enterprise Linux 5.9
    Red Hat certified
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    Red Hat Enterprise Linux 6.0 – 6.3
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are distributed through a download available at
    http://www.microsoft.com/en-us/download/details.aspx?id=34603
    Red Hat Enterprise Linux 6.4
    Red Hat certified
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    SUSE Linux Enterprise Server 11 with Service Pack 2 and Service Pack 3
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    Open SUSE 12.1
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    Ubuntu 12.04 and 12.10
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    Ubuntu 13.04 and 13.10
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    Oracle Linux 6.4
    Drivers: IDE, SCSI, networking, and mouse
    Services: Time synchronization, operating system shutdown, and heartbeat
    Integration services are built-in and do not require a separate download and installation.
    Note
    Oracle Linux is only supported when running the Red Hat Compatible Kernel.
    ==

  • Physical processors limit in Oracle Linux 5

    Hello:
    the Linux administrator told me that Oracle Linux 5 allows up to 4 physical processors.
    I've searched the Internet. I've found in http://www.oracle.com/us/technologies/linux/025994.htm that the allowed number of CPUs for a Oracle Linux 5 OS is unlimited. I've send this link to the Linux Admin, but his answer was that this specification refers only to logical CPUs, and that the limitation has to do with the number of physical processors.
    I've searched again the Internet, but I could not find anything about this restriction. Then, my question is: is threre any limit on the number of physical processors on Oracle Linux 5 OS? Where is it documented?
    Thank you and best regards,

    I've opened a SR (3-2552873331) to the Oracle Support. Here is their answer:
    Hi Marco,
    Yes you can download and install the Oracle Linux on the mentioned Box. The CPU limitations Vs price will come in picture only while registering for support.
    Thank you for your answer.
    Does it mean that threre is no limit on CPU sockets other than the imposed by the RedHat kernel (64/255 logical CPUs for Redhat 5)? Is it possible to download Oracle Linux 5 from the network and install it on a i.e. 8 dualcore box, setting the support license aside (no support is needed)? Do I need to change and recompile the linux kernel of the Oracle Linux 5 to allow the installation on a 8 dualcore box?
    Regards,
    Marco Antonio Serrano
    Oracle Linux and Oracle VM is free to download and use on any platform configuration for any purpose. Oracle charges only for support.
    Please visit the product pricing page at http://www.oracle.com/us/corporate/pricing/price-lists/index.html for the complete details, but in essence, Oracle licenses   support based on the number of CPU sockets (chips) and not by cores (virtual processors). Premier support levels permit unlimited numbers of CPU sockets, otherwise only a maximum of 2 cores are allowed.
    + Sincerely, +
    Hello:
    the Linux administrator told me that Oracle Linux 5 allows up to 4 physical processors.
    I've searched the Internet. I've found in http://www.oracle.com/us/technologies/linux/025994.htm that the allowed number of CPUs for a Oracle Linux 5 OS is unlimited. I've send this link to the Linux Admin, but his answer was that this specification refers only to logical CPUs, and that the limitation has to do with the number of physical processors.
    I've searched again the Internet, but I could not find anything about this restriction. Then, my question is: is threre any limit on the number of physical processors on Oracle Linux 5 OS? Where is it documented?
    Thank you and best regards,
    Regards,
    Marco A. Serrano

  • Dedicate more CPU space to Virtual Machine

    Hi,
    I have an VM with 8 virtual cores dedicated to it. The physical VM has 8 cores and it's a Xeon CPU E5-2650.
    However, when I am doing a stress test on the VM, the VM is getting really slow (normal of course) and it says CPU usage 100%.
    But, when I check the physical CPU usage (hyper-V server 2012 core - Task Manager) it says well, maximum 4% usage.. and all my other VM's are still performing very good.
    Is there a way to dedicate a much bigger part to a VM? I really need that option.
    Thanks in advance!

    I have responded to this quite a bit in this forum over the years.  (do some searches of the forum, you will find a lot of information).
    1) how are you stressing the CPU?  If your CPU test is single threaded it will not perform well in a VM, no matter the hypervisor.
    The CPU use in the VM is what the VM experiences, not the impact on the host.
    The nuts and bolts are discussed there:
    http://social.technet.microsoft.com/wiki/contents/articles/1234.hyper-v-concepts-vcpu-virtual-processor.aspx
    What your VM is doing and how your application behaves (uses the processor) is highly relevant to the discussion.  As well as how you evaluate it.
    Increasing weight allows your VM process thread to cut in line of other process threads, giving more CPU time, but having a negative impact on other VMs.  Reserve will have no impact until your hypervisor runs out of CPU processing power.  Affinity
    also does not impact CPU scheduling.
    Giving your Hyper-V Server a faster CPU will always give you a win.
    Brian Ehlert
    http://ITProctology.blogspot.com
    Learn. Apply. Repeat.
    Disclaimer: Attempting change is of your own free will.

Maybe you are looking for