[Gluster-devel] cache doesn t seem to be taken into account

[Sebastien COUPPEY]

On Thu, Jan 10, 2008 at 01:44:46AM +0530, Anand Avati wrote:
> can you try the same test without AFR?

umh without AFR, I have good results :

couppey:/tmp# time cp /mnt/speedTest/128Mo.img ./

real					0m11.215s
user					0m0.012s
sys						0m0.396s
couppey:/tmp# time cp /mnt/speedTest/128Mo.img ./

real					0m2.060s
user					0m0.012s
sys						0m0.464s
couppey:/tmp# time cp /mnt/speedTest/128Mo.img ./

real					0m4.771s
user					0m0.012s
sys						0m0.476s
couppey:/tmp# time cp /mnt/speedTest/128Mo.img ./

real					0m4.515s
user					0m0.012s
sys						0m0.476s
couppey:/tmp# time cp /mnt/speedTest/128Mo.img ./

real					0m11.399s
user					0m0.016s
sys						0m0.428s
couppey:/tmp# time cp /mnt/speedTest/128Mo.img ./

real					0m4.177s
user					0m0.008s
sys						0m0.496s
couppey:/tmp# time cp /mnt/speedTest/128Mo.img ./

real					0m2.002s
user					0m0.020s
sys						0m0.460s
couppey:/tmp# time cp /mnt/speedTest/128Mo.img ./

real					0m2.413s
user					0m0.004s
sys						0m0.496s



>  the reason might be inconsistant
> mtime being returned by AFR which is being discussed in another thread
> between Krishna and Li Daobing. Also, we suggest you upgrade to
> glusterfs--mainline--2.5--patch-628 from the TLA.

unfortunatelly after 2 h I am not able to compile it on RH5 or Centos
5, (tla is not even packaged for these systems) even with adding all the libs fuse, flex .... 
python2.4 present on the system is not recognise.

I think I ll have to wait for the .deb and .rpm of the 1.3.8

> 
> > I discovered the GLFS 2 days ago, and have been since testing it.
> 
> 
> Good luck with the testing! do reach us for questions.
> 
> avati