problem with nfs ls latency during high IO
mike at cchtml.com
Tue Mar 15 14:52:17 UTC 2011
Judith Flo Gaya on 03/15/2011 09:23 AM wrote:
> I browsed the code, but I can't see from what I read that the change was
> applied on my current working version (2.6.35) and the rest as they look
> very similar to me.
> If that is the case and the patch is applied, then why is the behavior
> so different between red hat and fedora? Why has the code changed so
> much between the two kernels?
I wouldn't know the exact answer to that question, but the kernel
changes very frequently. RHEL5 is 17 versions behind F14. RHEL6 is
already 3 behind F14. If you want to be sure that the NFS code is all
identical then you should ask on the kernel list or an NFS list.
> It already works fine with red hat 6, my issue is with fedora.
> The longest hang for red hat had been 40s, with fedora I have minutes of
> delay. I know I can buy Red had subscriptions but I need the fast
> updates from fedora.
There might be different kernel defaults for RHEL6. I have not studied
the two distributions that closely.
However, 40 seconds does not seem ideal to me either. Have you
considered it may be an issue with your IBRIX? I'm not familiar with
those devices to make a suggestion though. I ran a simple test with 4.5
gig file and continuously listed the destination directory from the
client during the transfer. It took at most 5 seconds to see the listing
out of 6 listing attempts. I was getting about 100mb/sec throughput
(1gbit network). Fedora 14 server and Fedora 14 client.
> I'm sorry if my questions seems to be rather stupid, I'm not very
> skilled with the kernel-code related issues.
There are no stupid questions. :)
More information about the users