Hi,
This night, while half a-sleep I thought it is usefull to have a tool
which creates a number of files in a directory and then starts to randomly
rename them. During this, it should output how much it has done and how
many renames per second it did.
5 minutes back I programmed such a program, you can download it from:
http://www.vanheusden.com/Linux/rename_test-1.0.tgz
But now, fully awake with at least 8 cups of coffee in my system I cannot
think of anything usefull this program is actually doing.
Well, maybe to test if something gets corrupted allong the way?
Oh well.
Folkert van Heusden
http://www.vanheusden.com
On Jan 20, 2003 18:29 +0100, [email protected] wrote:
> This night, while half a-sleep I thought it is usefull to have a tool
> which creates a number of files in a directory and then starts to randomly
> rename them. During this, it should output how much it has done and how
> many renames per second it did.
> 5 minutes back I programmed such a program, you can download it from:
> http://www.vanheusden.com/Linux/rename_test-1.0.tgz
>
> But now, fully awake with at least 8 cups of coffee in my system I cannot
> think of anything usefull this program is actually doing.
> Well, maybe to test if something gets corrupted allong the way?
For sure correctness is important (we use all sorts of benchmarks like
this to try and exercise our filesystem code), but also performance
measurement is important too. For example, mailservers do lots of file
renames to push mail through various stages in the delivery pipeline.
Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/
> Hi,
>
> This night, while half a-sleep I thought it is usefull to have a tool
> which creates a number of files in a directory and then starts to randomly
> rename them. During this, it should output how much it has done and how
> many renames per second it did.
> 5 minutes back I programmed such a program, you can download it from:
> http://www.vanheusden.com/Linux/rename_test-1.0.tgz
>
> But now, fully awake with at least 8 cups of coffee in my system I cannot
> think of anything usefull this program is actually doing.
> Well, maybe to test if something gets corrupted allong the way?
>
> Oh well.
Here's a suggestion:
Take one Very Old Kernel (say, 2.4.2)
and one Very New Kernel (2.5.59 ??)
Run your test on both, and see what differences you see.
If you see a change greater than 10%, that's a sign your test is
useful...or not.
cliffw
>
>
> Folkert van Heusden
> http://www.vanheusden.com
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
> have a tool
> which creates a number of files in a directory and then starts to randomly
> rename them. During this, it should output how much it has done and how
> many renames per second it did.
> 5 minutes back I programmed such a program, you can download it from:
> http://www.vanheusden.com/Linux/rename_test-1.0.tgz
CW> Here's a suggestion:
CW> Take one Very Old Kernel (say, 2.4.2)
CW> and one Very New Kernel (2.5.59 ??)
CW> Run your test on both, and see what differences you see.
CW> If you see a change greater than 10%, that's a sign your test is
CW> useful...or not.
Ok, I do not have a system with a 2.5 kernel, but here are some results
anyway:
after 210.000 renames:
2.2.20 ext2 1535/s
2.4.20 ext2 3087/s !!
ext3 1263/s !?
On Mon, 20 Jan 2003 [email protected] wrote:
> This night, while half a-sleep I thought it is usefull to have a tool
> which creates a number of files in a directory and then starts to
> randomly rename them.
> But now, fully awake with at least 8 cups of coffee in my system I cannot
> think of anything usefull this program is actually doing.
> Well, maybe to test if something gets corrupted allong the way?
You could test with NFS, you might even break something...
Rik
--
Bravely reimplemented by the knights who say "NIH".
http://www.surriel.com/ http://guru.conectiva.com/
Current spamtrap: <a href=mailto:"[email protected]">[email protected]</a>
Rik van Riel writes:
> On Mon, 20 Jan 2003 [email protected] wrote:
>
> > This night, while half a-sleep I thought it is usefull to have a tool
> > which creates a number of files in a directory and then starts to
> > randomly rename them.
>
> > But now, fully awake with at least 8 cups of coffee in my system I cannot
> > think of anything usefull this program is actually doing.
> > Well, maybe to test if something gets corrupted allong the way?
>
> You could test with NFS, you might even break something...
Now, when NFS has been mentioned.
To test reiser{fs,4} we use
ftp://ftp.namesys.com/pub/namesys-utils/nfs_fh_stale.c which creates
given number of threads each performing randomly reads, writes,
truncates, renames, fsyncs, and unlinks over the shared set of files. As
its name implies it was initially created to test support for the stale
NFS handle detection, but turned out to be quite useful for the general
testing also.
>
> Rik
Nikita.