Re: Bug#775014: nfs-common: Degraded performance on nfs4 clients after upgrade to Jessie
I am no debian kernel maintainer, but as I have seen this, some suggestions to
try:
Am Samstag, 10. Januar 2015, 01:59:50 schrieb Leopold Palomo-Avellaneda:
> Package: nfs-common
> Version: 1:1.2.8-9
> Severity: important
>
> I have a typical environment with a nfs server and some clients with home
> there, or shared resources. In the server I have a Wheezy (in one server
> with 3.12-0.bpo.1-amd64 version and in other 3.2.63). My clients basically
> are Wheezy but I'm preparing to Jessie and I wanted to test the upgrade.
I suggest you upgrade to 3.16 bpo kernel. Maybe that already makes a
difference. And additionally there is greater chance you get security updates
on that one, cause AFAIK older bpo kernels are not maintained anymore.
> I have realized that in the two cases that I have tested (at home and work)
> both, with the same parameters (fstab) that worked "normal" with Wheezy,
> now, they have an horrible performance in write operations. I'm having
> differences of several order of magnitude (37,9 KB/s Jessie vs 17,4 MB/s)
>
> For instance:
> Wheezy machine:
> $ dd if=/dev/zero of=myrandom bs=1024 count=1000
[…]
> 1024000 octets (1,0 MB) copiats, 0,0590127 s, 17,4 MB/s
>
> Jessie machine
> $ dd if=/dev/zero of=myrandom bs=1024 count=1000
[…]
> 1024000 octets (1,0 MB) copiats, 27,0021 s, 37,9 kB/s
>
> I hope that the kernel developers have changed some default parameter and I
> have it wrong. But, with this version of the kernel, my nfs clients cannot
> work well.
>
> I'm open to make any test or provide any needed information.
>
> -- Package-specific info:
[…]
> -- /etc/fstab --
> #nfs4
> cami:/recursos /home/recursos
> nfs4
auto,rw,nodev,sync,_netdev,retry=10,rsize=65536,wsize=65536,soft,noat
> ime,intr 0 0 -- /proc/mounts --
> cami:/recursos /home/recursos nfs4
> rw,sync,nodev,noatime,vers=4.0,rsize=65536,wsize=65536,namlen=255,soft,prot
> o=tcp,port=0,timeo=600,retrans=2,sec=sys,clientaddr=192.168.1.8,local_lock=n
> one,addr=192.168.1.254 0 0
What is this on the wheezy machine?
Look in grep nfs /proc/mounts or nfsstat -m
I suggest not to manually tune this. Current kernels use good default values.
I have seen rsize and wsize of up to 1048576.
Additionally are they all the same hardware, same network cards? If not, what
is the fast system using and what is the slow system using?
> -- System Information:
> Debian Release: 8.0
> APT prefers testing
> APT policy: (500, 'testing'), (1, 'unstable')
> Architecture: amd64 (x86_64)
> Foreign Architectures: i386
>
> Kernel: Linux 3.16.0-4-amd64 (SMP w/4 CPU cores)
[…]
--
Martin 'Helios' Steigerwald - http://www.Lichtvoll.de
GPG: 03B0 0D6C 0040 0710 4AFA B82F 991B EAAC A599 84C7
Reply to: