[Users] RE: slow fsync rate

[Marc Aymerich]

On Wed, Aug 25, 2010 at 8:54 AM, Dietmar Maurer <dietmar at proxmox.com> wrote:

> Hi Kir,
>
> > These numbers very much resemble fsync() rate with write cache enabled
> > (~1000/sec) and disabled (50-70/sec).
> > check write cache settings with hdparm + check whether you have barrier
> > mount option on ext3.
>
> I use exactly the same system. The only thing I change is the kernel.
> kernel with
> OpenVZ patches is factor 20 slower!
>
> That is why I ask if you can reproduce the problem?
>

poor performance here Dietmar, whit a SAS 10Krpm drive:

m2:~# sysbench --test=fileio --file-num=1 --file-total-size=20G
--file-fsync-all=on --file-test-mode=seqrewr --max-time=100
--file-block-size=4096 --max-requests=0 run
sysbench 0.4.12:  multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 1

Extra file open flags: 0
1 files, 20Gb each
20Gb total file size
Block size 4Kb
Calling fsync() after each write operation.
Using synchronous I/O mode
Doing sequential rewrite test
Threads started!
Time limit exceeded, exiting...
Done.

Operations performed:  0 Read, 3323 Write, 3323 Other = 6646 Total
Read 0b  Written 12.98Mb  Total transferred 12.98Mb  (132.88Kb/sec)
   33.22 Requests/sec executed

Test execution summary:
    total time:                          100.0286s
    total number of events:              3323
    total time taken by event execution: 100.0146
    per-request statistics:
         min:                                 24.22ms
         avg:                                 30.10ms
         max:                                 89.84ms
         approx.  95 percentile:              33.97ms

Threads fairness:
    events (avg/stddev):           3323.0000/0.00
    execution time (avg/stddev):   100.0146/0.00

m2:~# uname -a
Linux m2 2.6.32.15 #1 SMP Fri Jul 16 20:17:30 CEST 2010 x86_64 GNU/Linux

I think it's budarin.1 kernel.



-- 
Marc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://openvz.org/pipermail/users/attachments/20100826/50cf19f9/attachment-0001.html