NFS mount points don't appear in disk monitoring
NFS mount points don't appear in disk monitoring
hobbit 4.2.0 running on RHEL 4, one monitored box is Sun sparc box(Solaris 9) with hobbit client installed.
local disks look like correct, but NFS mount points don't appear in disk monitoring:
bash-2.05$ df -k Filesystem kbytes used avail capacity Mounted on /dev/dsk/c0t0d0s0 4950381 3163019 1737859 65% / /proc 0 0 0 0% /proc mnttab 0 0 0 0% /etc/mnttab fd 0 0 0 0% /dev/fd swap 1788448 112 1788336 1% /var/run swap 1788360 24 1788336 1% /tmp /dev/dsk/c0t3d0s0 8705501 5090262 3528184 60% /ora /dev/dsk/c0t2d0s0 8705501 2909068 5709378 34% /oracle /dev/dsk/c0t0d0s7 2703017 1989310 659647 76% /export/home
10.1.1.1:/vol/fv_RmanOraData1 12582912 170284 12412628 2% /oradata1 10.1.1.1:/vol/fv_RmanOraData 12582912 6496568 6086344 52% /oradata 10.1.1.1:/vol/fv_RmanOraLog 6291456 36192 6255264 1% /oralog
only local disks are displayed in disk monitoring:
Tue May 22 20:30:23 CST 2007 - Filesystems OK
Filesystem kbytes used avail capacity Mounted on swap 1789312 24 1789288 1% /tmp /dev/dsk/c0t0d0s0 4950381 3163021 1737857 65% / /dev/dsk/c0t3d0s0 8705501 5090262 3528184 60% /ora /dev/dsk/c0t2d0s0 8705501 2909068 5709378 34% /oracle /dev/dsk/c0t0d0s7 2703017 1989312 659645 76% /export/home
I even added following section before "DEFAULT" section in hobbit-clients.cfg file, but no effect.
HOST=sun01 DISK / 90 95 DISK /oralog 90 95 DISK %^/ora.* 90 95
Thanks very much.
On Wednesday 23 May 2007, case wrote:
NFS mount points don't appear in disk monitoring For me, this is a feature.
You can change this if you open the hobbitclient-linux.sh on the client. Find the df section and change the df command untill it sends all the file system information you need.
Stef
On Wed, May 23, 2007 at 04:05:10PM +0800, case wrote:
NFS mount points don't appear in disk monitoring
This is by design - checking NFS filesystems can cause the Hobbit client to stall indefinitely if the nfs system becomes unresponsive.
I'd recommend checking nfs filesystems from the server they are exported from. It's also a lot more effective, since you will only get 1 alarm when it starts to fill up, instead of all 200 hosts that have it mounted triggering disk alerts.
Regards, Henrik
participants (3)
-
casedj@21cn.com
-
henrik@hswn.dk
-
stef.coene@docum.org