Re: nfs write problem & "handle_workitem_freeblocks: block count"

[Jan Niemann <jan.niemann@tu-bs.de>]

hello david, hi all,

On Sun, Jun 01, 2003 at 12:40:20AM -0700, David S. wrote:
> > after some frustrating hours and reading pr's 1580 and 3210:
> > is anybody out there able to _write_ to an nfs-share ?
> > if you can, how do you do it ?
> 
> No problem:
> 
> 	doctorno:9% echo 'Hello, world!' > ./tmp0

what about writing bigger chunks of data (>5mb,<2gb) ?
dd'ing /dev/zero for a while ?

> 	doctorno:13% df
> 	Filesystem         1K-blocks     Used    Avail Capacity  Mounted on
> 	/dev/wd0a             126639    29295    91013    24%    /
> 	/dev/wd0e            8885338  2292525  6148547    27%    /usr
> 	/dev/wd0d             126671    10227   110111     8%    /var
> 	mfs:13488             507759       12   482360     0%    /tmp
> 	kernfs                261740      372        0   100%    /kern
> 	procfs                 81296    50672        0   100%    /proc
> 	amd:381                    0        0        0   100%    /home
> 	amd:381                    0        0        0   100%    /mnt
> 	goldfinger:/vol/20  17144376  2467006 13820150    15%    /a/goldfinger/vol/20
> 	doctorno:14% uname -a
> 	OpenBSD doctorno 3.3 GENERIC#0 i386

what are your mount-options ?
i can't imagine that using amd would help me...


here's my setup, the shares in question are b and d :

jan@hamlet[jan] /bin/df|grep "vol\|^F"
Filesystem            512-blocks     Used    Avail Capacity  Mounted on
haegar:/volumes/a      233588534  3348628 218560480     2%    /nfs/a
haegar:/volumes/c      137651644 36791046 93978016    28%    /nfs/c
haegar:/volumes/b      233588534 112903392 109005716    51%    /nfs/b
haegar:/volumes/d      137651644 95548120 35220942    73%    /nfs/d
jan@hamlet[jan] /sbin/mount|grep vol
haegar:/volumes/a on /nfs/a type nfs (nodev, noexec, nosuid, read-only, v3, udp, timeo=100)
haegar:/volumes/c on /nfs/c type nfs (nodev, noexec, nosuid, read-only, v3, udp, timeo=100)
haegar:/volumes/b on /nfs/b type nfs (nodev, noexec, nosuid, v2, udp, timeo=100)
haegar:/volumes/d on /nfs/d type nfs (nodev, noexec, nosuid, v3, udp, timeo=100)

are the filesystems to big for the client ?

jan