questions about speed of Read/Write through mergerfs(FUSE based) system

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

questions about speed of Read/Write through mergerfs(FUSE based) system

Pahome Chen
I want to test the Read/Write speed of a raid0(4 x ssd ) through mergerfs.
1. I mount radi0 on ~/a .Then I mount ~/a in ~/mnt by mergerfs.
2. I use fio to test one file in ~/mnt/test so that means ~/a/test exists. 
3. Here is my library version:
  • FUSE library vresion 2.9.7
  • fusermount version 2.9.7
  • FUSE kernel interface version 7.19
  • Ubuntu 4.4.0-38 x86_64
  • fio_version = 2.14

Below is my fio config file

[global]
ioengine=libaio
direct=0
time_based
runtime=60
ramp_time=30
size=64g
group_reporting

[S_100RW_1M_R]
rw=read
numjobs=1
iodepth=32
bs=1m
stonewall

[S_100RW_1M_W]
rw=write
numjobs=1
iodepth=32
bs=1m
stonewall

when I execute :sudo fio fio.cfg and it done
it shows the result of seq. read and seq. write.

I expected Read and Write are almost 1200(MB/s) cuz of 4 ssd.

And read is 1200MB/s ( it's great cuz I use 4 ssd), but Write is 420MB/s.

If I test the raid0 not through fuse, Read/Write speed are all 1200MB/s up

so I wonder why write performance is so bad after through fuse?

It's a normal case when I go through fuse-based file systems?

Did anyone test the disk performance through mergerfs or fuse-based fs?

thanks


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
--
fuse-devel mailing list
To unsubscribe or subscribe, visit https://lists.sourceforge.net/lists/listinfo/fuse-devel