On 02/03/06, David Laight <david@l8s.co.uk> wrote:
> On Thu, Mar 02, 2006 at 05:25:51PM +1000, Ray Phillips wrote:
> >
> > I used a 75 GB target file on a ~ 108 GB /usr partition and ls -l
> > /usr confirms it is 75 GB in size.  However, df says /usr has 59 GB
> > available, which doesn't add up.
>
> What does 'ls -ls' show, the file could be sparse

(question the same here)

pcsupp2$\ ls -ls /data/iscsi-target0
14727744 -rw-r--r--  1 root  wheel  52428800001 Dec 16 02:31 /data/iscsi-ta=
rget0

pcsupp2$\ df -k /data
Filesystem  1K-blocks      Used     Avail Capacity  Mounted on
/dev/wd0j    53592164  14741124  36171432    28%    /data

I am also having some weird performance problems with iSCSI on this
target - the idea was to try to utilize an old 200Mhz P/Pro box for
this. The initiator is W2003 server (latest 2.01 version from
Microsoft, the initiator software that is...). The read speed appears
fine - 100Mb/s cards on a switch gives some 20Mb/s measured by ntop;
writing starts fast enough, but gradually grinds down to about 2Mb/s
and even less. The CPU load at the target stays at about 8%. The
memory is just 128MB, so it may be some paging effects:

pcsupp2$\ TERM=3D top
top: no termcap entry for a `' terminal
load averages:  1.58,  1.31,  0.76    up 5 days,  1:39    up 5 days,=20
1:39   02:34:05
47 processes:  46 sleeping, 1 on processor
CPU states:     % user,     % nice,     % system,     % interrupt,     % id=
le
Memory: 60M Act, 32M Inact, 4260K Wired, 5692K Exec, 66M File, 5200K Free
Swap: 384M Total, 104M Used, 280M Free

  PID USERNAME PRI NICE   SIZE   RES STATE      TIME   WCPU    CPU COMMAND
19151 nobody     2    0    19M   22M select    45:04  7.37%  7.37% ntop
17328 root       2    0    83M  692K netio      3:01  3.47%  3.47% <iscsi-t=
arge
    8 root      18    0     0K   19M syncer     3:34  0.15%  0.15% [ioflush=
]
    9 root     -18    0     0K   19M aiodoned   1:14  0.00%  0.00% [aiodone=
d]
  105 root       2    0  1176K 1324K select     0:41  0.00%  0.00% sendmail
  282 root       2    0   148K  612K select     0:40  0.00%  0.00% ypbind
    7 root     -18    0     0K   19M pgdaemon   0:30  0.00%  0.00% [pagedae=
mon]
  822 root      10    0   232K  336K nanoslee   0:03  0.00%  0.00% cron
  273 root       2    0   328K  228K poll       0:02  0.00%  0.00% rpcbind
  235 root       2    0   180K  192K kqread     0:02  0.00%  0.00% syslogd
  692 root       2    0   292K    4K select     0:02  0.00%  0.00% <sshd>
    1 root      10    0    64K  204K wait       0:01  0.00%  0.00% init
 1057 ci         2    0   380K  680K select     0:01  0.00%  0.00% sshd
  412 root       2    0   176K  196K select     0:01  0.00%  0.00% rpc.lock=
d
  400 root       2    0   172K  192K select     0:01  0.00%  0.00% rpc.stat=
d
  760 ci         2    0   380K    4K select     0:01  0.00%  0.00% <sshd>
    0 root     -18    0     0K   19M schedule   0:01  0.00%  0.00% [swapper=
]
10990 root      28    0   152K  900K CPU        0:00  0.00%  0.00% top



>
>         David
>
> --
> David Laight: david@l8s.co.uk
>

Chavdar