Take the 2-minute tour ×
Unix & Linux Stack Exchange is a question and answer site for users of Linux, FreeBSD and other Un*x-like operating systems.. It's 100% free, no registration required.

Ok.

I get about 1/3 of the speed on the same drives when using ZFS.

Now, I understand that ZFS is self-healing and that everything is also stored with its hashes. This makes it safer to use.

I haven't used raidz yet, so this excludes that part of the equation - but please tell me if by using raidz I can speed things up.

If the hashes are computed for comparison - where and when does this happen.

Where does the speed issue come into play?

I'm happy now that I have mirrored and safe data, on my system, but would like to know how I can tweak performance.

share|improve this question
2  
What system are you using? If Linux it is because it goes through FUSE so it goes from userspace to kernel back to another piece of userspece back to kernel. –  Maciej Piechotka Jan 3 '11 at 13:39
    
You get about 1/3 of the speed as … what? Your question is meaningless unless you describe the systems you're comparing: what filesystem are you comparing against? What operating system, running on what hardware, configured how? What are you measuring (reading/writing small/large files/directories etc.)? –  Gilles Jan 3 '11 at 19:30
    
@Gilles: To be more specific I have 10+ drives running JFS (previously on Gentoo). The initial conversion of JFS to ZFS on Gentoo + Fuse gave me about 1/3 of the speed. These are unmirrored (until after conversion), non-raid drives. I'm talking about 5-7 megabytes per second (JFS into new ZFS) vs a normal drive to drive (JFS to JFS copy) of 25-29 megabytes per second. This is regardless of Fuse on Gentoo or FreeBSD access. The data insertion seems to be about 1/3 of the speed. Other (read) access is much more respectable. –  Forkrul Assail Jan 4 '11 at 11:48
    
Did you try ZFS on Linux via the kernel? –  CMCDragonkai Jul 8 at 10:43
add comment

1 Answer 1

up vote 2 down vote accepted

In case of FUSE and ZFS, performance is awful, and that's understandable.

FreeBSD tuning guide gives some pointers on this. Article is long, and there is many specific points for different architectures, so I won't copy it to here. Most important point is that you need lots of memory, one gigabyte being minimum.

raidz only helps if disk IO performance is your bottleneck. If it's software bugs, processor or memory, it's not going to be better with that.

share|improve this answer
    
Thanks, I'll look into that. –  Forkrul Assail Feb 22 '11 at 6:52
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.