FreeBSD 7.1 disk performance issue on ESXi 3.5

[Ivan Voras]

Sebastiaan van Erk wrote:
> Sebastiaan van Erk wrote:
>> (However, just to give you an idea I attached the basic 5.1.2
>> unixbench outputs (the CPU info for FreeBSD is "fake", since unixbench
>> does a cat /proc/cpuinfo, so I removed the /proc/ part and copied the
>> output under linux to the "procinfo" file.)


   System: test-fbsd.vpn1.sebster.com: FreeBSD
------------------------------------------------------------------------
Benchmark Run: Tue Feb 10 2009 06:25:49 - 06:54:08
2 CPUs in system; running 1 parallel copy of tests

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   14144383.9   1212.0
Double-Precision Whetstone                       55.0       3238.7    588.9
Execl Throughput                                 43.0        630.0    146.5
File Copy 1024 bufsize 2000 maxblocks          3960.0      28793.2     72.7
File Copy 256 bufsize 500 maxblocks            1655.0      33410.0    201.9
File Copy 4096 bufsize 8000 maxblocks          5800.0      33536.8     57.8
Pipe Throughput                               12440.0    1146784.7    921.9
Pipe-based Context Switching                   4000.0      36203.6     90.5
Process Creation                                126.0        783.3     62.2
Shell Scripts (1 concurrent)                     42.4        645.1    152.2
Shell Scripts (8 concurrent)                      6.0        115.4    192.3
System Call Overhead                          15000.0     939647.5    626.4
                                                                   ========
System Benchmarks Index Score                                         212.4


   System: test-ubuntu: GNU/Linux
------------------------------------------------------------------------
Benchmark Run: Mon Feb 09 2009 15:15:06 - 15:43:20
2 CPUs in system; running 1 parallel copy of tests

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   18610575.3   1594.7
Double-Precision Whetstone                       55.0       2990.1    543.7
Execl Throughput                                 43.0       1058.6    246.2
File Copy 1024 bufsize 2000 maxblocks          3960.0     468973.2   1184.3
File Copy 256 bufsize 500 maxblocks            1655.0     132022.2    797.7
File Copy 4096 bufsize 8000 maxblocks          5800.0     921448.5   1588.7
Pipe Throughput                               12440.0    1132933.6    910.7
Pipe-based Context Switching                   4000.0      93429.0    233.6
Process Creation                                126.0       1744.3    138.4
Shell Scripts (1 concurrent)                     42.4       2566.9    605.4
Shell Scripts (8 concurrent)                      6.0        518.4    864.0
System Call Overhead                          15000.0    1935577.0   1290.4
                                                                   ========
System Benchmarks Index Score                                         656.1


The results are ... interesting. It seems that FreeBSD simply dies in
any test having a high context switch rate. Hmmm, this looks familiar.
Either I or a collegue of mine had a similar situation some time ago,
with the same discrepancy in disk speeds and the same difference in
context switches. Unfortunately, there was no solution.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
Url : http://lists.freebsd.org/pipermail/freebsd-performance/attachments/20090210/46dfb036/signature.pgp