Hobbit, Net-snmp snmptrap,and HPOV
All,
I'm trying to push alarms from Hobbit to HPOV. We have a custom script that is invoking Netsnmp's snmptrap. The snmp traps are reaching the HP Node Manager Event Browser correctly. However, the OV operations trap condition is not recognizing the trap. The trap is being displayed as "unmatched" in the Message Browser even though there is a condition set up for that specific trap.
When we look at the unmatched trap, it appears that the OID is not being recognized and an '0' is inserted, for example instead of .1.3.6.1.4.1.7058.0.XX, we are seeing .0.0.XX.
Any guidance would be greatly appreaciated.
Thanks,
-Mark Coupe
Coupe, Mark (ITD) wrote:
All,
I'm trying to push alarms from Hobbit to HPOV. We have a custom script that is invoking Netsnmp's snmptrap. The snmp traps are reaching the HP Node Manager Event Browser correctly. However, the OV operations trap condition is not recognizing the trap. The trap is being displayed as "unmatched" in the Message Browser even though there is a condition set up for that specific trap.
When we look at the unmatched trap, it appears that the OID is not being recognized and an '0' is inserted, for example instead of .1.3.6.1.4.1.7058.0.XX, we are seeing .0.0.XX.
Any guidance would be greatly appreaciated.
Thanks,
-Mark Coupe
I, and perhaps others would be interested in what precisely you are doing? Are you pushing all hobbit alarms from the hobbit server to HPOV?
Sorry, I am HPOV ignorant, but we are interested in forwarding all the hobbit alarms/alerts as SNMP traps to one of Concord's products. I am not to familiar with Concord's product either. I just know the powers that be would like to have snmp traps from all our monitored hosts.
In otherwords, I am interested in any system or methodology that someone has setup to forward alarms from the hobbit server to an SNMP server. Any setups out there that someone is willing to share the details?
David
David Gore wrote:
Coupe, Mark (ITD) wrote:
All,
I'm trying to push alarms from Hobbit to HPOV. We have a custom script that is invoking Netsnmp's snmptrap. The snmp traps are reaching the HP Node Manager Event Browser correctly. However, the OV operations trap condition is not recognizing the trap. The trap is being displayed as "unmatched" in the Message Browser even though there is a condition set up for that specific trap.
When we look at the unmatched trap, it appears that the OID is not being recognized and an '0' is inserted, for example instead of .1.3.6.1.4.1.7058.0.XX, we are seeing .0.0.XX.
Any guidance would be greatly appreaciated.
Thanks,
-Mark Coupe
I, and perhaps others would be interested in what precisely you are doing? Are you pushing all hobbit alarms from the hobbit server to HPOV?
Sorry, I am HPOV ignorant, but we are interested in forwarding all the hobbit alarms/alerts as SNMP traps to one of Concord's products. I am not to familiar with Concord's product either. I just know the powers that be would like to have snmp traps from all our monitored hosts.
In otherwords, I am interested in any system or methodology that someone has setup to forward alarms from the hobbit server to an SNMP server. Any setups out there that someone is willing to share the details?
David
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
Here is Mark's script posted with permission:
SCRIPT
#!/bin/bash
#SVCCODES="disk:100,cpu:200,procs:300,svcs:350,msgs:400,conn:500,http:60 0,ftp:721,ssh:722,ssh1:722,ssh2:722,telnet:723,smtp:725,fping:727,mq:729 ,dns:800,pop2:809,pop-2:809,pop:810,pop3:810,pop-3:810,nntp:819,imap:843 ,imap2:843,imap3:843,imap4:843,mrtg:850,swap:860,memory:870,sys:880,err: 890,test:901"
#Variables #$OID - OID to be used without trailing value (.1.3.6.1.4.1.7058)(BB default) #$OIDTAIL - ending value of OID particluar to alarm type; determined by "case" statement #$MACHIP - IP address of device generating alarm; determined by hobbit #$BBHOSTNAME - hostname of device generating alarm; determined by the hobbit bb-hosts file #$BBSVCNUM - service code corresponding to the service being monitored; determined by hobbit configuration #$BBCOLORLEVEL - alarm level; determined by hobbit #$RECOVERED - '0' denotes alarm; '1' denotes recovery #$HPOV - the IP address of the destination HPOV server #TMPDIR - directory for location of temporary file necessary for sed of $MACHIP
###Start script ###
CHRONO=date
OID=.1.3.6.1.4.1.7058
#Put your own OV server IP address in the next line.
HPOV=xxx.xxx.xxx.xxx
TMPDIR=/home/hobbit/server/ext
case $BBSVCNUM in 100) #Disk OIDTAIL=10 ;; 200) #CPU OIDTAIL=20 ;; 300) #Procs OIDTAIL=30 ;; 350) #Svcs OIDTAIL=40 ;; 400) #Msgs OIDTAIL=50 ;; 500) #Conn OIDTAIL=60 ;; 600) #Http OIDTAIL=70 ;; #OIDTAIL=80 (not assigned) 721) #FTP OIDTAIL=90 ;; 722) #SSH OIDTAIL=100 ;; 723) #Telnet OIDTAIL=110 ;; 725) #SMTP OIDTAIL=120 ;; 727) #Fping OIDTAIL=130 ;; 729) #MQ OIDTAIL=140 ;; 800) #DNS OIDTAIL=150 ;; 809) #POP2 OIDTAIL=160 ;; 810) #POP OIDTAIL=170 ;; 819) #NNTP OIDTAIL=180 ;; 843) #IMAP OIDTAIL=190 ;; 850) #MRTG OIDTAIL=200 ;; 860) #Swap OIDTAIL=210 ;; 870) #Memory OIDTAIL=220 ;; 880) #Sys OIDTAIL=230 ;; 890) #ERR OIDTAIL=240 ;; 901) #Test OIDTAIL=250 ;; esac
if [[ "$RECOVERED" = "0" ]]; then if [[ "$BBCOLORLEVEL" = "purple" ]]; then OIDTAIL=$(($OIDTAIL + 1)) elif [[ "$BBCOLORLEVEL" = "yellow" ]]; then OIDTAIL=$(($OIDTAIL + 2)) elif [[ "$BBCOLORLEVEL" = "red" ]]; then OIDTAIL=$(($OIDTAIL + 3)) fi elif [[ "$RECOVERED" = "1" ]]; then BBCOLORLEVEL=green fi
#Following line: reads Hobbit IP into temporary file echo $MACHIP > $TMPDIR/tmpfile.txt
#Following lines: puts Hobbit 12 digit IP format in dotted decimal and
removes any leading zeroes.
MACHIP=sed 's/\(...\)\(...\)\(...\)\(...\)/\1.\2.\3.\4/' $TMPDIR/tmpfile.txt
echo $MACHIP > $TMPDIR/tmpfile.txt
MACHIP=sed -e 's/^00//' -e 's/^0//' -e 's/\.000/./' -e 's/\.00/./' -e 's/\.0/./' -e 's/\.000/./' -e 's/\.00/./' -e 's/\.0/./' -e 's/\.000/./' -e 's/\.00/./' -e 's/\.0/./' -e 's/\.\.\./.0.0./' -e 's/\.\./.0./' -e 's/\.$/.0/' $TMPDIR/tmpfile.txt
#Following line: removes the temporary file rm -f $TMPDIR/tmpfile.txt
#Following line: forwards the error message to HPOV as an SNMP message
snmptrap -v 1 -c mitcovo $HPOV "$OID" "$MACHIP" 6 $OIDTAIL ""
echo "DATE: $CHRONO" >> /home/hobbit/server/ext/hpov-push.log echo "snmptrap -v 1 -c public $HPOV "$OID" "$MACHIP" 6 $OIDTAIL "" " >> /home/hobbit/server/ext/hpov-push.log
END SCRIPT
On Fri, Sep 02, 2005 at 11:43:55AM -0400, Coupe, Mark (ITD) wrote:
I'm trying to push alarms from Hobbit to HPOV. We have a custom script that is invoking Netsnmp's snmptrap. The snmp traps are reaching the HP Node Manager Event Browser correctly. However, the OV operations trap condition is not recognizing the trap. The trap is being displayed as "unmatched" in the Message Browser even though there is a condition set up for that specific trap.
When we look at the unmatched trap, it appears that the OID is not being recognized and an '0' is inserted, for example instead of .1.3.6.1.4.1.7058.0.XX, we are seeing .0.0.XX.
The problem is somewhere in your custom script, since that's what generates the OID that is being fed to the snmptrap command.
The command you should end up with ought to look somewhat like
snmptrap -v1 -c public
172.16.10.100
""
hobbit.foo.com
0 0 ""
1.3.6.1.4.1.7058.1
s "This is a Hobbit trap message"
172.16.10.100 is the HPOV monitor IP-address, "hobbit.foo.com" is the hostname reporting the trap, "1.3.6...." is the trap OID.
At least, this generates an SNMP trap that Ethereal decodes as I would expect.
Regards, Henrik
I hope this is not another silly question! I've searched the archives, but not seen anything that solves my problem, which may prove to be a misconfiguration. I have a bb-dftab on my clients in /home/hobbit/client/etc/bb-dftab. After changing it, I restarted the hobbit client just to be sure. /home and /opt are still showing up as yellow and red respectively. /home (90 %) has reached the WARNING level (90 %) /opt (97 %) has reached the PANIC level (95 %)
The client bb-dftab contains: /usr:92:98 /home:95:99 /opt:99:100 and is owned by the hobbit user, with rw-r--r-- permissions
The client hobbitclient.cfg is default as bundled and contains:
For the disk check
DODISK="TRUE" NODISKCOLOR="clear" BBDFTAB="$BBHOME/etc/bb-dftab" DFEXCLUDE="cdrom" DFPANIC="95" DFSORT="4" DFUSE="^/dev" DFWARN="90"
participants (4)
-
Allan.Marillier@dana.com
-
David.Gore@mci.com
-
henrik@hswn.dk
-
Mark.Coupe@state.ma.us