Virtual SCSI performance

From: Heikki Manninen (hma_at_ee.oulu.fi)
Date: 04/26/05 

Date: Tue, 26 Apr 2005 10:16:59 +0000 (UTC)

I'm having a performance issue with a small micro
partitioning/virtual I/O test system.

Hardware:

  9113-550
  4 x Power5 1,5GHz, 8GB RAM, 2 x 36GB internal disks
  5703 PCI-X Ultra320
  2104-DS4 + 4 x 4073 73GB 10k disks

Software:

  Partition 1)

    Virtual I/O Server
    512MB RAM, 1.0 cpu, 4 virtual cpus
    2 x 8GB mirrored LV on 2104-DS4, exported through virtual scsi

  Partition 2)

    RHEL3 AS
    1GB RAM, 2.0 cpu, 8 virtual cpus
    2 x 8GB virtual scsi disks from partition 1

When observing disk throughput (topas) on partition 1 during
I/O operations like creating an LV copy, I get a steady
24MB/s read and 24MB/s write rates for each hdisk.

However when doing disk I/O testing on partition 2 and at the same
time observing throughput again on vio server the read and write rates
are steadily around 4MB/s for each hdisk. VIO server cpu utilization is
around 12% so that should not be the issue.

Any ideas how to improve performance or what could be the cause? 

-- 
Heikki M

Relevant Pages

  • IO WAIT Information From IBM
    ... the I/O wait metric in AIX. ... AIX scheduler, the CPU "queues", the CPU states, and the idle or wait ... To summarize it in one sentence, 'iowait' is the percentage ...
    (AIX-L)
  • userland starvation with 2.4.25-rc2
    ... I am trying to load the SCSI disk by doing in parallel: ... The upper graph is the CPU load, the lower graph is the I/O load. ... you can see, at certain points the kernel will take all available CPU, ...
    (Linux-Kernel)
  • Re: IO WAIT Information From IBM
    ... >understanding of how the I/O wait value is collected and calculated. ... >of the CPU resource. ... The wait processes only job is to increment the counters that report ... >Each CPU can be in one of four states: user, sys, idle, iowait. ...
    (AIX-L)
  • Re: IO WAIT Information From IBM
    ... >understanding of how the I/O wait value is collected and calculated. ... >of the CPU resource. ... The wait processes only job is to increment the counters that report ... >Each CPU can be in one of four states: user, sys, idle, iowait. ...
    (AIX-L)
  • Re: It is almost certain now, INTEL will have 64bit x86 !!
    ... >>between this an 8400 which was that you had CPU boards, memory ... > if the MaxCPU boards had memory) number of CPUs. ... I/O bridge x 16! ... In the F15K these can be used for I/O or 2 CPU MAXCPU cards ...
    (comp.os.vms)
Quantcast