On Friday, 28 July 2000 at 19:21:10 +0400, Maxim Konovalov wrote:
>
> Hello,
>
> I've got a question about vinum. There are two scsi disks with vinum
> stripe. Here is a vinum config:
>
> drive a device /dev/da2s1e
> drive b device /dev/da3s1e
> volume vinum0
> plex name vinum0.p0 org striped 1024s vol vinum0
> sd name vinum0.p0.s0 drive a plex vinum0.p0
> len 34791424s driveoffset 265s plexoffset 0s
> sd name vinum0.p0.s1 drive b plex vinum0.p0
> len 34791424s driveoffset 265s plexoffset 1024s
Please show the output of 'vinum list'.
> # uname -sr
> FreeBSD 4.1-STABLE
>
> There are not any other fs on these disks. The problem is - when I run
> iozone (iozone 4096 /logs/io0.tmp) I get a very strange result:
>
> # iostat -d da1 da3 10
> da1 da3
> KB/t tps MB/s KB/t tps MB/s
> 0.00 0 0.00 0.00 0 0.00
> 4.50 107 0.47 0.00 0 0.00
> 4.51 98 0.43 0.00 0 0.00
> 4.50 116 0.51 0.00 0 0.00
> 4.49 117 0.52 0.00 0 0.00
> 4.50 111 0.49 0.00 0 0.00
> ....
>
> You see, there is not activity on da3. Are there any explanations?
I think this says more about iozone than it does about Vinum. Try
running rawio instead and see what happens.
Note also that a stripe size which is a power of two is not a good
idea. It can give rise to the kind of behaviour you report: since
cylinder groups are usually 32 MB in size, this arrangement will put
all the super blocks on the same subdisk. Use an odd number, say 283
kB, instead.
Greg
--
Finger [EMAIL PROTECTED] for PGP public key
See complete headers for address and phone numbers
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
