Date: Mon, 2 Sep 2013 10:10:18 -0400 (EDT) From: Rick Macklem <rmacklem@uoguelph.ca> To: "Sam Fourman Jr." <sfourman@gmail.com> Cc: FreeBSD FS <freebsd-fs@freebsd.org> Subject: Re: NFS on ZFS pure SSD pool Message-ID: <590302855.16994708.1378131018328.JavaMail.root@uoguelph.ca> In-Reply-To: <CAOFF%2BZ2C26V3Ktuvj5RFwGN%2BH657OPXm796TqA54xOpQj1TVFg@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Sam Fourman Jr. wrote:
[lots of stuff snipped for brevity]
> root@students:/users # nfsstat -e -s
>
> Server Info:
> Getattr Setattr Lookup Readlink Read Write Create
> Remove
> 106273793 1417764 19593633 12021 2497674 7927757 1047249
> 772450
> Rename Link Symlink Mkdir Rmdir Readdir RdirPlus
> Access
> 319284 924 13813 63500 20980 526257 0
> 677005862
I didn't spot this when it first was posted, because of the wrap around.
That's a *lot* of Access operations, imho. Maybe tweaking the Mac OS X client
could reduce these?
I see there are a couple of options in their nfs.conf(5) file that might help?
nfs.client.access_cache_timeout
nfs.client.access_for_getattr
You could also look at their mount_nfs man page to see if there are other
settings for access related stuff.
As well, I'm wondering if the Macs may be doing Access ops like crazy because
they see that ACLs are enabled.
I think ZFS always have ACLs enabled, but you can change the line in
sys/fs/nfs/nfs_commonsubs.c
int nfsrv_useacl = 1;
to
int nfsrv_useacl = 0;
and then build/boot a new kernel on the server to disable them. (It isn't
a sysctl, because it normally depends on the server file system to say if
they are supported.)
Also, pull up a terminal window and do an "ls -l" on some directory, to make
sure everything isn;t owned by "nobody". If it is, the name/uid mapping for
NFSv4 isn;t working correctly.
Good luck with it, rick
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?590302855.16994708.1378131018328.JavaMail.root>