Xymon Mailing List Archive search

Help with disk monitoring

4 messages in this thread

list Edward Bailey · Mon, 11 Dec 2006 16:21:16 -0500 ·
Hello,

I have recently inherited a sizable Hobbit environment so please have
patience with my general ignorance of all things Hobbit. Our Hobbit
works really well except for one notable exception and I can't figure
out why. We have a series of boxes that run an application and for some
reason these systems will not monitor the /opt file system and alert us
when /opt goes over 90% even though these boxes have the same Hobbit
setup like the others on the server side and the client side.

The only difference I can see is that the /opt file system is mapped
differently on these systems. Most systems use something like

/dev/sda5 120G 82G 73% /opt

But the problem systems use

/dev/mapper/sysvg-opt 120G   82G   32G  73% /opt


Also the file systems from these problem servers show up differently as
well. The standard file system mounts are displayed without an issues
but the LVM maps display the mount point but nothing else

/dev/mapper/sysvg-opt
/dev/mapper/sysvg-var
/dev/cciss/c0d0p1      2016016    216060   1697544  12% /
/dev/cciss/c0d0p2      3020172    950024   1916728  34% /usr


The problem servers also have the follow error in the client side log

Mon Dec 11 15:53:21 2006 bb-local.sh /opt/Big_Brother/bb/bin/bb-disk.sh:
line 180: test: : integer expression expected
Mon Dec 11 15:58:21 2006 bb-local.sh /opt/Big_Brother/bb/bin/bb-disk.sh:
line 180: test: : integer expression expected
Mon Dec 11 15:58:21 2006 bb-local.sh /opt/Big_Brother/bb/bin/bb-disk.sh:
line 180: test: : integer expression expected
Mon Dec 11 16:03:21 2006 bb-local.sh /opt/Big_Brother/bb/bin/bb-disk.sh:
line 180: test: : integer expression expected

From bbdef.sh

#
# LOCAL CLIENT MONITORING CONFIGURATION FOR bb-local.sh
#
# WARNING AND PANIC LEVELS FOR LOCAL SYSTEM INFOMRAION
# YOU CAN SET VALUES ON A SPECIFIC FILESYSTEM BY USING
#       THE etc/bb-dftab FILE
#

DFWARN=90                               # (YELLOW) DISK % TO WARN
DFPANIC=95                              # (RED) DISK % TO PANIC
export DFWARN DFPANIC

The lines above and below line 180 of bb-disk.sh

 MARK=""
                        if test "$DISKPCT" -ge "$DFWARNVAL"         #
ABOVE WARNING LEVEL
                        then
                                if test "$DISKPCT" -ge "$DFPANICVAL"
# ABOVE PANIC, RED ALERT !!!
                                then
                                        COLOR="red"
                                        MARK="PANIC"
                                        MARKLEVEL=$DFPANICVAL
                                        HIGHMARK="PANIC"
                                        GIFCOLOR="&red"
                                        REDLINE="${GIFCOLOR} ${DISKNAME}
(${DISKPCT}%) has reached the defined disk space ${MARK} le
vel (${MARKLEVEL}%)
${REDLINE}"
                                else


By default all servers are set to alert at 90% and 95% percent.

Thanks in advance for everyone's help

Ed Bailey

CONFIDENTIALITY NOTICE
This e-mail and any attachments contain information which may be confidential or privileged and exempt from disclosure under applicable law.  If you are not the intended recipient, be aware that any disclosure, copying, distribution, or use of the contents of this information is without authorization and is prohibited.  If you have received this email in error, please immediately notify us by returning it to the sender and delete this copy from your computer system.  Thank you.
list Henrik Størner · Mon, 11 Dec 2006 22:27:37 +0100 ·
quoted from Edward Bailey
On Mon, Dec 11, 2006 at 04:21:16PM -0500, Bailey, Edward wrote:
Also the file systems from these problem servers show up differently as
well. The standard file system mounts are displayed without an issues
but the LVM maps display the mount point but nothing else

/dev/mapper/sysvg-opt
/dev/mapper/sysvg-var
/dev/cciss/c0d0p1      2016016    216060   1697544  12% /
/dev/cciss/c0d0p2      3020172    950024   1916728  34% /usr
This is a classical problem with the Big Brother client - you are
apparently using the BB client instead of the Hobbit client. What's
happening is that the "df" command used on the monitored servers breaks
long lines in the output, so your "df" output looks like

Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/sda1               579598    192382    356293  36% /
/dev/mapper/vg1-lv_usr
                       2568144    648964   1788724  27% /usr
/dev/mapper/vg1-lv_var
                       1048540    733448    315092  70% /var

and since the BB client does a 'grep "^/dev"' to weed out entries like
tmpfs, you only get the /dev/mapper line, not the next line with the 
actual data in it.

The solution is to change the DFCMD setting on your servers to
"df -P", making it use the POSIX format output.


Regards,
Henrik
list Edward Bailey · Mon, 11 Dec 2006 16:42:01 -0500 ·
Thanks for your quick response:) I did see that we have the big brother
client installed so I need to plan on upgrading the clients first.

Thanks again

Ed 
quoted from Henrik Størner
-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] Sent: Monday, December 11, 2006 4:28 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] Help with disk monitoring

On Mon, Dec 11, 2006 at 04:21:16PM -0500, Bailey, Edward wrote:
Also the file systems from these problem servers show up differently > as well. The standard file system mounts are displayed without an > issues but the LVM maps display the mount point but nothing else
/dev/mapper/sysvg-opt
/dev/mapper/sysvg-var
/dev/cciss/c0d0p1      2016016    216060   1697544  12% /
/dev/cciss/c0d0p2      3020172    950024   1916728  34% /usr
This is a classical problem with the Big Brother client - you are apparently using the BB client instead of the Hobbit client. What's happening is that the "df" command used on the monitored servers breaks long lines in the output, so your "df" output looks like

Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/sda1               579598    192382    356293  36% /
/dev/mapper/vg1-lv_usr
                       2568144    648964   1788724  27% /usr
/dev/mapper/vg1-lv_var
                       1048540    733448    315092  70% /var

and since the BB client does a 'grep "^/dev"' to weed out entries like tmpfs, you only get the /dev/mapper line, not the next line with the actual data in it.

The solution is to change the DFCMD setting on your servers to "df -P", making it use the POSIX format output.


Regards,
Henrik

CONFIDENTIALITY NOTICE
This e-mail and any attachments contain information which may be confidential or privileged and exempt from disclosure under applicable law.  If you are not the intended recipient, be aware that any disclosure, copying, distribution, or use of the contents of this information is without authorization and is prohibited.  If you have received this email in error, please immediately notify us by returning it to the sender and delete this copy from your computer system.  Thank you.
list Pnixon · Mon, 11 Dec 2006 16:46:08 -0500 ·
In the meantime, before you get around to all your clients,
 you can edit the bbsys.local file and change

DF="/bin/df -k"
Line to:
DF="/bin/df -kP"

Which will take out the extra line return.

*PS - Dangit Henrik! I knew this answer and was looking for it to see where
to change it! Hehehe!
quoted from Edward Bailey
 
-----Original Message-----
From: Bailey, Edward [mailto:user-dcede94722d2@xymon.invalid] Sent: Monday, December 11, 2006 4:42 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: RE: [hobbit] Help with disk monitoring

Thanks for your quick response:) I did see that we have the big brother
client installed so I need to plan on upgrading the clients first.

Thanks again

Ed 
-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid]
Sent: Monday, December 11, 2006 4:28 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] Help with disk monitoring

On Mon, Dec 11, 2006 at 04:21:16PM -0500, Bailey, Edward wrote:
Also the file systems from these problem servers show up
differently
as well. The standard file system mounts are displayed without an > issues but the LVM maps display the mount point but nothing else
/dev/mapper/sysvg-opt
/dev/mapper/sysvg-var
/dev/cciss/c0d0p1      2016016    216060   1697544  12% /
/dev/cciss/c0d0p2      3020172    950024   1916728  34% /usr
This is a classical problem with the Big Brother client - you are apparently using the BB client instead of the Hobbit client. What's happening is that the "df" command used on the monitored servers breaks long lines in the output, so your "df" output looks like

Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/sda1               579598    192382    356293  36% /
/dev/mapper/vg1-lv_usr
                       2568144    648964   1788724  27% /usr
/dev/mapper/vg1-lv_var
                       1048540    733448    315092  70% /var

and since the BB client does a 'grep "^/dev"' to weed out entries like tmpfs, you only get the /dev/mapper line, not the next line with the actual data in it.

The solution is to change the DFCMD setting on your servers to "df -P", making it use the POSIX format output.


Regards,
Henrik

CONFIDENTIALITY NOTICE
This e-mail and any attachments contain information which may be
confidential or privileged and exempt from disclosure under applicable law.
If you are not the intended recipient, be aware that any disclosure,
copying, distribution, or use of the contents of this information is without
authorization and is prohibited.  If you have received this email in error,
please immediately notify us by returning it to the sender and delete this
copy from your computer system.  Thank you.