Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


UnixBench running 4 parallel processes scores lower then one on Xen PV?
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

UnixBench running 4 parallel processes scores lower then one on Xen PV?

bitronictechbitronictech Member
edited November 2012 in Help

I am running a Xen PV on a Quad-Core 3.3GHz E3-1230-V2 Xeon Ivy Bridge, 8GB RAM, 2 X 1TB SAS with 8 threads. One thread is pinned to Dom0. I had setup a VPS with 1024MB RAM and 50GB HD. Capped at 100% of one core but not pinned, with 4vCPUs. DomU is running CentOS 6.3, Dom0 uses virtualizor as the frontend panel. When I ran benchmarks I got some... interesting results:

# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #

Version 5.1.3 Based on the Byte Magazine Unix Benchmark

Multi-CPU version Version 5 revisions by Ian Smith,
Sunnyvale, CA, USA
January 13, 2011 johantheghost at yahoo period com

1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10

1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10

1 x Execl Throughput 1 2 3

1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3

1 x File Copy 256 bufsize 500 maxblocks 1 2 3

1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3

1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10

1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10

1 x Process Creation 1 2 3

1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10

1 x Shell Scripts (1 concurrent) 1 2 3

1 x Shell Scripts (8 concurrent) 1 2 3

4 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10

4 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10

4 x Execl Throughput 1 2 3

4 x File Copy 1024 bufsize 2000 maxblocks 1 2 3

4 x File Copy 256 bufsize 500 maxblocks 1 2 3

4 x File Copy 4096 bufsize 8000 maxblocks 1 2 3

4 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10

4 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10

4 x Process Creation 1 2 3

4 x System Call Overhead 1 2 3 4 5 6 7 8 9 10

4 x Shell Scripts (1 concurrent) 1 2 3

4 x Shell Scripts (8 concurrent) 1 2 3

========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)

System: testraid.bitronictech.net: GNU/Linux
OS: GNU/Linux -- 2.6.32-279.el6.x86_64 -- #1 SMP Fri Jun 22 12:19:21 UTC 2012
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz (6585.0 bogomips)
x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET, Intel virtualization
CPU 1: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz (6585.0 bogomips)
x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET, Intel virtualization
CPU 2: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz (6585.0 bogomips)
x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET, Intel virtualization
CPU 3: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz (6585.0 bogomips)
x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET, Intel virtualization
01:46:04 up 54 min, 1 user, load average: 0.17, 0.04, 0.17; runlevel 3


Benchmark Run: Thu Nov 15 2012 01:46:04 - 02:14:17
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 36081254.9 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4007.1 MWIPS (9.8 s, 7 samples)
Execl Throughput 671.9 lps (29.5 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 293866.3 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 76780.8 KBps (30.1 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 940053.1 KBps (30.0 s, 2 samples)
Pipe Throughput 429204.8 lps (10.0 s, 7 samples)
Pipe-based Context Switching 75300.2 lps (10.0 s, 7 samples)
Process Creation 1281.5 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 906.8 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 148.1 lpm (60.2 s, 2 samples)
System Call Overhead 394699.7 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 36081254.9 3091.8
Double-Precision Whetstone 55.0 4007.1 728.6
Execl Throughput 43.0 671.9 156.3
File Copy 1024 bufsize 2000 maxblocks 3960.0 293866.3 742.1
File Copy 256 bufsize 500 maxblocks 1655.0 76780.8 463.9
File Copy 4096 bufsize 8000 maxblocks 5800.0 940053.1 1620.8
Pipe Throughput 12440.0 429204.8 345.0
Pipe-based Context Switching 4000.0 75300.2 188.3
Process Creation 126.0 1281.5 101.7
Shell Scripts (1 concurrent) 42.4 906.8 213.9
Shell Scripts (8 concurrent) 6.0 148.1 246.8
System Call Overhead 15000.0 394699.7 263.1
========
System Benchmarks Index Score 402.4


Benchmark Run: Thu Nov 15 2012 02:14:17 - 02:42:47
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 19209588.0 lps (10.1 s, 7 samples)
Double-Precision Whetstone 2115.9 MWIPS (9.0 s, 7 samples)
Execl Throughput 649.2 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 76680.8 KBps (30.1 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 20547.6 KBps (30.1 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 268725.1 KBps (30.1 s, 2 samples)
Pipe Throughput 228735.3 lps (10.0 s, 7 samples)
Pipe-based Context Switching 38611.6 lps (10.1 s, 7 samples)
Process Creation 1314.8 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1195.1 lpm (60.2 s, 2 samples)
Shell Scripts (8 concurrent) 173.6 lpm (60.6 s, 2 samples)
System Call Overhead 218916.0 lps (10.1 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 19209588.0 1646.1
Double-Precision Whetstone 55.0 2115.9 384.7
Execl Throughput 43.0 649.2 151.0
File Copy 1024 bufsize 2000 maxblocks 3960.0 76680.8 193.6
File Copy 256 bufsize 500 maxblocks 1655.0 20547.6 124.2
File Copy 4096 bufsize 8000 maxblocks 5800.0 268725.1 463.3
Pipe Throughput 12440.0 228735.3 183.9
Pipe-based Context Switching 4000.0 38611.6 96.5
Process Creation 126.0 1314.8 104.3
Shell Scripts (1 concurrent) 42.4 1195.1 281.9
Shell Scripts (8 concurrent) 6.0 173.6 289.4
System Call Overhead 15000.0 218916.0 145.9
========
System Benchmarks Index Score 231.3

Comments

  • I/O Seek Test (No Cache)

    ioping -RD
    463 iops, 1.8 mb/s
    min/avg/max/mdev = 0.1/2.2/149.1/5.3 ms
    I/O Reads - Sequential
    ioping -RL
    707 iops, 176.7 mb/s
    min/avg/max/mdev = 0.7/1.4/67.9/3.5 ms
    I/O Reads - Cached
    ioping -RC
    184762 iops, 721.7 mb/s
    min/avg/max/mdev = 0.0/0.0/0.0/0.0 ms

  • DD

    dd if=/dev/zero of=sb-io-test bs=1M count=1k conv=fdatasync
    13.7568 s, 78.1 MB/s
    dd if=/dev/zero of=sb-io-test bs=64k count=16k conv=fdatasync
    15.9546 s, 67.3 MB/s
    dd if=/dev/zero of=sb-io-test bs=1M count=1k oflag=dsync
    67.8212 s, 15.8 MB/s
    dd if=/dev/zero of=sb-io-test bs=64k count=16k oflag=dsync
    773.597 s, 1.4 MB/s

    FIO

    Read IOPS 914.0
    Read Bandwidth 3.6 MB/second
    Write IOPS 631.0
    Write Bandwidth 2.5 MB/second
    Raw FIO Output
    FIO random reads:
    randomreads: (g=0): rw=randread, bs=4K-4K/4K-4K, ioengine=libaio, iodepth=64
    fio-2.0.9
    Starting 1 process
    randomreads: Laying out IO file(s) (1 file(s) / 1024MB)

    randomreads: (groupid=0, jobs=1): err= 0: pid=1320: Thu Nov 15 01:20:20 2012
    read : io=1024.3MB, bw=3659.1KB/s, iops=914 , runt=286567msec
    cpu : usr=0.78%, sys=2.51%, ctx=257244, majf=0, minf=89
    IO depths : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=0.1%, 32=0.1%, >=64=100.0%
    submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
    complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.1%, >=64=0.0%
    issued : total=r=262207/w=0/d=0, short=r=0/w=0/d=0

    Run status group 0 (all jobs):
    READ: io=1024.3MB, aggrb=3659KB/s, minb=3659KB/s, maxb=3659KB/s, mint=286567msec, maxt=286567msec

    Disk stats (read/write):
    xvda1: ios=262136/18, merge=68/6, ticks=18349234/15444, in_queue=18367665, util=100.00%
    Done

    FIO random writes:
    randomwrites: (g=0): rw=randwrite, bs=4K-4K/4K-4K, ioengine=libaio, iodepth=64
    fio-2.0.9
    Starting 1 process

    randomwrites: (groupid=0, jobs=1): err= 0: pid=1325: Thu Nov 15 01:27:15 2012
    write: io=1024.3MB, bw=2527.4KB/s, iops=631 , runt=414992msec
    cpu : usr=0.37%, sys=1.97%, ctx=257645, majf=0, minf=25
    IO depths : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=0.1%, 32=0.1%, >=64=100.0%
    submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
    complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.1%, >=64=0.0%
    issued : total=r=0/w=262207/d=0, short=r=0/w=0/d=0

    Run status group 0 (all jobs):
    WRITE: io=1024.3MB, aggrb=2527KB/s, minb=2527KB/s, maxb=2527KB/s, mint=414992msec, maxt=414992msec

    Disk stats (read/write):
    xvda1: ios=0/262295, merge=0/131, ticks=0/26718024, in_queue=26727750, util=100.00%
    Done`

    What is the cause of this madness? Any ideas?

  • I have seen interesting UnixBench results myself with Xen PV. Can you run a ServerBear.com benchmark and post the link please? Thank you.

  • http://serverbear.com/benchmark/2012/11/15/PKHFMvRW12XsR2KH

    The output above was from that benchmark.

    Any input would be greatly appreciated.

  • @bitronictech - try benching two VPS servers at the same time on the same node. Xen allocates CPU power in relation to other VMs. I have seen similar issues with Xen, and I am even maintaining my own Xen stack for CentOS 6. On that CPU you should be getting way more with 4 cores. Did you change the weights for Dom0 and DomU? What OS and Xen version are you running?

  • This is my XM dmesg

    xm dmesg


    \ \/ /___ _ __
    \ // _ \ \047_ \
    / \ / | | |
    /_/__
    |_| |_|

    _____ _ ____ _____ ___ ___ _ __ _ _ ____
    |___ / / | |___ \ |___ / / _ \ ( _ ) / |/ /_ / | | | ___|
    |
    \ | | __) |
    |_ | | | |/ _ \ | | \047_ \ | | / _ \ |___ \
    ) || | / __/||) | || | () || | () || || __/ |) |
    |____()()_____| |____/ ___/ ___()|___()()__||____/

    http://www.cl.cam.ac.uk/netos/xen
    University of Cambridge Computer Laboratory

    Xen version 3.1.2-308.16.1.el5 ([email protected]) (gcc version 4.1.2 20080704 (Red Hat 4.1.2-52)) Tue Oct 2 21:53:43 EDT 2012
    Latest ChangeSet: unavailable

    (XEN) Command line: dom0_mem=2048M
    (XEN) Video information:
    (XEN) VGA is text mode 80x25, font 8x16
    (XEN) VBE/DDC methods: V2; EDID transfer time: 1 seconds
    (XEN) Disc information:
    (XEN) Found 2 MBR signatures
    (XEN) Found 2 EDD information structures
    (XEN) Xen-e820 RAM map:
    (XEN) 0000000000000000 - 000000000009d400 (usable)
    (XEN) 000000000009d400 - 00000000000a0000 (reserved)
    (XEN) 00000000000e0000 - 0000000000100000 (reserved)
    (XEN) 0000000000100000 - 00000000bc036000 (usable)
    (XEN) 00000000bc036000 - 00000000bc038000 (reserved)
    (XEN) 00000000bc038000 - 00000000bf5bd000 (usable)
    (XEN) 00000000bf5bd000 - 00000000bf5cf000 (reserved)
    (XEN) 00000000bf5cf000 - 00000000bf5d0000 (usable)
    (XEN) 00000000bf5d0000 - 00000000bf5e1000 (reserved)
    (XEN) 00000000bf5e1000 - 00000000bf5f2000 (usable)
    (XEN) 00000000bf5f2000 - 00000000bf612000 (reserved)
    (XEN) 00000000bf612000 - 00000000bf63f000 (usable)
    (XEN) 00000000bf63f000 - 00000000bf6bf000 (reserved)
    (XEN) 00000000bf6bf000 - 00000000bf7bf000 (ACPI NVS)
    (XEN) 00000000bf7bf000 - 00000000bf7f0000 (ACPI data)
    (XEN) 00000000bf7f0000 - 00000000bf800000 (usable)
    (XEN) 00000000bf800000 - 00000000c0000000 (reserved)
    (XEN) 00000000e0000000 - 00000000f0000000 (reserved)
    (XEN) 00000000feb00000 - 00000000feb04000 (reserved)
    (XEN) 00000000fec00000 - 00000000fec01000 (reserved)
    (XEN) 00000000fed10000 - 00000000fed1a000 (reserved)
    (XEN) 00000000fed1c000 - 00000000fed20000 (reserved)
    (XEN) 00000000fee00000 - 00000000fee01000 (reserved)
    (XEN) 00000000ffc00000 - 0000000100000000 (reserved)
    (XEN) 0000000100000000 - 0000000840000000 (usable)
    (XEN) System RAM: 32757MB (33543836kB)
    (XEN) Xen heap: 13MB (13932kB)
    (XEN) Domain heap initialised: DMA width 32 bits
    (XEN) Processor #0 7:10 APIC version 21
    (XEN) Processor #1 7:10 APIC version 21
    (XEN) Processor #2 7:10 APIC version 21
    (XEN) Processor #3 7:10 APIC version 21
    (XEN) Processor #4 7:10 APIC version 21
    (XEN) Processor #5 7:10 APIC version 21
    (XEN) Processor #6 7:10 APIC version 21
    (XEN) Processor #7 7:10 APIC version 21
    (XEN) IOAPIC[0]: apic_id 0, version 32, address 0xfec00000, GSI 0-23
    (XEN) Enabling APIC mode: Flat. Using 1 I/O APICs
    (XEN) Using scheduler: SMP Credit Scheduler (credit)
    (XEN) Detected 3292.614 MHz processor.
    (XEN)
    (XEN) VMX: Supported advanced features:
    (XEN) - APIC MMIO access virtualisation
    (XEN) - APIC TPR shadow
    (XEN) - Extended Page Tables (EPT)
    (XEN) - Virtual-Processor Identifiers (VPID)
    (XEN) - Virtual NMI
    (XEN) - MSR direct-access bitmap
    (XEN) VMX: VPID is available.
    (XEN) HVM: VMX enabled
    (XEN) HVM: Hardware Assisted Paging detected and enabled.
    (XEN) VMX: MSR intercept bitmap enabled
    (XEN) I/O virtualisation disabled
    (XEN) CPU0: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz stepping 09
    (XEN) Booting processor 1/1 eip 7c000
    (XEN)
    (XEN) VMX: VPID is available.
    (XEN) CPU1: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz stepping 09
    (XEN) Booting processor 2/2 eip 7c000
    (XEN)
    (XEN) VMX: VPID is available.
    (XEN) CPU2: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz stepping 09
    (XEN) Booting processor 3/3 eip 7c000
    (XEN)
    (XEN) VMX: VPID is available.
    (XEN) CPU3: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz stepping 09
    (XEN) Booting processor 4/4 eip 7c000
    (XEN)
    (XEN) VMX: VPID is available.
    (XEN) CPU4: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz stepping 09
    (XEN) Booting processor 5/5 eip 7c000
    (XEN)
    (XEN) VMX: VPID is available.
    (XEN) CPU5: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz stepping 09
    (XEN) Booting processor 6/6 eip 7c000
    (XEN)
    (XEN) VMX: VPID is available.
    (XEN) CPU6: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz stepping 09
    (XEN) Booting processor 7/7 eip 7c000
    (XEN)
    (XEN) VMX: VPID is available.
    (XEN) CPU7: Intel(R) Xeon(R) CPU E3-1230 V2 @ 3.30GHz stepping 09
    (XEN) Total of 8 processors activated.
    (XEN) ENABLING IO-APIC IRQs
    (XEN) -> Using new ACK method
    (XEN) Platform timer overflows in 14998 jiffies.
    (XEN) Platform timer is 14.318MHz HPET
    (XEN) Brought up 8 CPUs
    (XEN) *** LOADING DOMAIN 0 ***
    (XEN) elf_parse_binary: phdr: paddr=0xffffffff80200000 memsz=0x303d98
    (XEN) elf_parse_binary: phdr: paddr=0xffffffff80503e00 memsz=0x251fb0
    (XEN) elf_parse_binary: phdr: paddr=0xffffffff80756000 memsz=0xc08
    (XEN) elf_parse_binary: phdr: paddr=0xffffffff80758000 memsz=0x19550c
    (XEN) elf_parse_binary: memory: 0xffffffff80200000 -> 0xffffffff808ed50c
    (XEN) elf_xen_parse_note: GUEST_OS = "linux"
    (XEN) elf_xen_parse_note: GUEST_VERSION = "2.6"
    (XEN) elf_xen_parse_note: XEN_VERSION = "xen-3.0"
    (XEN) elf_xen_parse_note: VIRT_BASE = 0xffffffff80000000
    (XEN) elf_xen_parse_note: PADDR_OFFSET = 0xffffffff80000000
    (XEN) elf_xen_parse_note: ENTRY = 0xffffffff80200000
    (XEN) elf_xen_parse_note: HYPERCALL_PAGE = 0xffffffff80206000
    (XEN) elf_xen_parse_note: FEATURES = "writable_page_tables|writable_descriptor_tables|auto_translated_physmap|pae_pgdir_above_4gb|supervisor_mode_kernel"
    (XEN) elf_xen_parse_note: LOADER = "generic"
    (XEN) elf_xen_addr_calc_check: addresses:
    (XEN) virt_base = 0xffffffff80000000
    (XEN) elf_paddr_offset = 0xffffffff80000000
    (XEN) virt_offset = 0x0
    (XEN) virt_kstart = 0xffffffff80200000
    (XEN) virt_kend = 0xffffffff808ed50c
    (XEN) virt_entry = 0xffffffff80200000
    (XEN) Xen kernel: 64-bit, lsb, compat32
    (XEN) Dom0 kernel: 64-bit, lsb, paddr 0xffffffff80200000 -> 0xffffffff808ed50c
    (XEN) PHYSICAL MEMORY ARRANGEMENT:
    (XEN) Dom0 alloc.: 0000000816000000->0000000818000000 (516096 pages to be allocated)
    (XEN) VIRTUAL MEMORY ARRANGEMENT:
    (XEN) Loaded kernel: ffffffff80200000->ffffffff808ed50c
    (XEN) Init. ramdisk: ffffffff808ee000->ffffffff81037000
    (XEN) Phys-Mach map: ffffffff81037000->ffffffff81437000
    (XEN) Start info: ffffffff81437000->ffffffff8143749c
    (XEN) Page tables: ffffffff81438000->ffffffff81447000
    (XEN) Boot stack: ffffffff81447000->ffffffff81448000
    (XEN) TOTAL: ffffffff80000000->ffffffff81800000
    (XEN) ENTRY ADDRESS: ffffffff80200000
    (XEN) Dom0 has maximum 8 VCPUs
    (XEN) elf_load_binary: phdr 0 at 0xffffffff80200000 -> 0xffffffff80503d98
    (XEN) elf_load_binary: phdr 1 at 0xffffffff80503e00 -> 0xffffffff80755db0
    (XEN) elf_load_binary: phdr 2 at 0xffffffff80756000 -> 0xffffffff80756c08
    (XEN) elf_load_binary: phdr 3 at 0xffffffff80758000 -> 0xffffffff8079aca8
    (XEN) Initrd len 0x749000, start at 0xffffffff808ee000
    (XEN) Scrubbing Free RAM:

  • .............................................................................................................................................................................................................................................................................................................done.
    (XEN) Xen trace buffers: disabled
    (XEN) Std. Loglevel: Errors and warnings
    (XEN) Guest Loglevel: Nothing (Rate-limited: Errors and warnings)
    (XEN) Xen is relinquishing VGA console.
    (XEN) *** Serial input -> DOM0 (type \047CTRL-a\047 three times to switch input to Xen).
    (XEN) Freed 116kB init memory.
    (XEN) save.c:232:d0 HVM restore: Xen changeset was not saved.
    (XEN) mm.c:630:d1 Non-privileged (1) attempt to map I/O space 00000000
    (XEN) mm.c:630:d1 Non-privileged (1) attempt to map I/O space 000000f0
    (XEN) mm.c:630:d1 Non-privileged (1) attempt to map I/O space 00000000
    (XEN) mm.c:630:d1 Non-privileged (1) attempt to map I/O space 000000f0
    (XEN) mm.c:630:d2 Non-privileged (2) attempt to map I/O space 00000000
    (XEN) mm.c:630:d2 Non-privileged (2) attempt to map I/O space 000000f0
    (XEN) mm.c:630:d2 Non-privileged (2) attempt to map I/O space 00000000
    (XEN) mm.c:630:d2 Non-privileged (2) attempt to map I/O space 000000f0

  • Dom0 is CentOS 5.8 Xen weight 512 with a core pinned and 2GB RAM
    DomU from that bench was a CentOS 6.3 with 1GB RAM weight 256 vCPUs CPU cap=100, so one whole thread, not pinned. There was plenty of free CPU as with the CPU cap it was only loading one thread.

  • @bitronictech try an leave DomU wide open, allocate 8 vCPUs (8 threads), no cpu cap and set a weight of 4096 for DomU. Run the benchmark again. If your scores don't increase, then something else is wrong there with Xen.

  • bitronictechbitronictech Member
    edited November 2012

    Dom0 has 8vCPUs rather, should it be pinned?

    Im going to run a benchmark now with the DomU settings you specified.

  • @bitronictech It doesn't really matter. The problem is with how Xen balances CPU resources.

  • @marcm If there is still an issue, is there a way to nail it down and correct it because this is really mystifying me! Thanks for all of the help thus far.

  • @bitronictech I am looking for a solution as well, and interestingly enough I am running Xen 4.1.3 on CentOS 6.3 with a custom 3.6 kernel. On my Xen E3 nodes it behaves right, but on my E5 nodes it shows similar issues. So do the benchmark and we'll go from there :)

  • bitronictechbitronictech Member
    edited November 2012

    It would seem my newbhood is showing here. The reason why the benchmark as so low on the second test was because it was simulating 4vCPUs going full tilt on one thread. I'll have to rethink my scheduling methods. So for that first package something like 100% CPU Cap on 1vCPU and for the next up 200% on 2vCPU and 400% on 4 vCPU. so on and so forth.

  • Weight is something to consider as well, I think round robin should work, but perhaps its better to give priority to higher paying customers.

  • May be is the Xen's CPU weight/priority yes.

  • Still a E3-1230v2 should give a better unixbench result even with 1 CPU core.

    My results with 4 CPU cores (although KVM virtualization) are:
    UnixBench (w/ all processors) 4650.6
    UnixBench (w/ one processor) 2100.1

    http://serverbear.com/benchmark/2012/10/5/zxuBhOL2po2edoai

  • @George_Fusioned said: Still a E3-1230v2 should give a better unixbench result even with 1 CPU core.

    @George_Fusioned - We offer both Xen and KVM, and I can tell you that with KVM you have no control over CPU allocation (using SolusVM of course), while with Xen you have more control such as weights, caps and so on. I'd say that the new results are in line with what I was expecting.

    @bitronictech - What did you change exactly in order to get those results?

  • @marcm That´s not correct. Use "cputune" a function of libvirt or cgroups. Both works without any problems.

  • marcmmarcm Member
    edited November 2012

    @fileMEDIA said: @marcm That´s not correct. Use "cputune" a function of libvirt or cgroups. Both works without any problems.

    @fileMEDIA - But I was talking about the options available in SolusVM. If one goes outside of what is available in SolusVM for KVM then you can tune a heck of allot more, like sockets, cores per socket and even threads per core (Hyperthreading) :)

  • You can activate cputune over custom config option. But correct, not all functions are integrated. A had done a feature request to phil for a including cputune feature. I hope it will come :)

  • @fileMEDIA said: You can activate cputune over custom config option. But correct, not all functions are integrated. A had done a feature request to phil for a including cputune feature. I hope it will come :)

    @fileMEDIA - Same here. I like KVM more and more over Xen. There is still a huge market for Xen, but KVM is a very cool technology. I hope that @soluslabs fixes the little swap initialization issue as well. It gets old after a while to tell customers to do swapoff, mkswap -> swapon - when using KVM templates. They don't like it, and I don't think that it is a good idea for us to redo every KVM template just to include a little script that does that. Other than that, I really like SolusVM :)

  • @marcm With a CPU cap of 100% in Xen, but with 4vCPUs it was trying to simulate 4 physical CPUs on one thread. I was able to get much better results with 1vCPU in the same parameters. That last benchmark I posted however was with all 8 cores wide open, however with only 1024 RAM.

  • @marcm Yeah, this bug isn´t nice.

    We switched from Xen to KVM. We have a few nodes with xen, but we migrate this to kvm very soon. We have more performance and more features with kvm and it performes like xen with few tweaks. I like the libvirt online migration function. I written a small php script to online migrate a vps without any shared storage. Only ./migrate.php id nodeid :)

  • I'm migrating from a server with a raid 0 array to a raid 10 to get rid of that nasty bottleneck with IO and increase redundancy. I have manually tar --numeric-owner every pv and restore them on the new node. Pain in the butt.

Sign In or Register to comment.