I don't like the idea of performing ping tests from the script so it just checks the status of the conn test, if it is red it scripts that host for the xsnmp round. It might be better to send a clear message. Hobbit will recheck the conn test if it is red anyway. My main problem was that we check 55 difference devices with bb-xsnmp and many of them for different tests like cpu, memory, temperature, uptime, fans, power, disk, bgp, etc. If an access point goes down, it is relatively minor for us but bb-xsnmp would hang causing about 300 purple alerts for good hosts. With this small modification only the one access point tests would go purple and everything else would be ok.
I am more of a C coder myself so I am sure someone else could improve on what I uploaded (that is mostly why I uploaded it ;) Butch

-----Original Message----- From: PNIXON at ci.somerville.ma.us [mailto:PNIXON at ci.somerville.ma.us] Sent: Tuesday, November 15, 2005 2:47 PM To: hobbit at hswn.dk Subject: RE: [hobbit] hobbit + bb-xsnmp.pl

Wouldn't it be better to query hobbit and check the status of the connection than run the ping test again? (Not that I could currently code this, but I've seen reference to this method before)

--Pat

-----Original Message----- From: Dan Vande More [mailto:bigdan at gmail.com] Sent: Tuesday, November 15, 2005 11:43 AM To: hobbit at hswn.dk Subject: Re: [hobbit] hobbit + bb-xsnmp.pl

The script was empty:) Thanks for the info though, I really like the idea of it pinging before checking. Also, I got hobbit to graph temperature by hacking it to spit out the format that hobbit likes. Note hobbit expects the third column to be degrees F, where as I just print the threshold. Since the thresholds aren't really acurate (Some say 750C, some say 65535C, etc), and hobbit doesn't use the F column, I just put it there for good measure.

Since a diff wouldn't do you any good either, here are the lines I've changed:

(line 2981, inside sub temperature_message) # compose message $message = "status $commahost.$test $worstcolor ".scalar(localtime)."\n\n"; my @indices = sort { $a <=> $b } keys(%statedb); foreach my $index (@indices) { my $description=$descrdb{$index}; $description=~s/ |:|-|\(|\)|\\|\//_/g; $description=~s/__/_/g; $description=lc($description); $message .= '&'.$colordb{$index}." $description\t".$valuedb{$index}."\t".$threshdb{$index}."\n"; } # foreach my $index (@indices)

The only thing I have to work on now is splitting the graphs. My 7609's have 47 temperature sensors apiece:)

Thanks again for the help.

Dan

On 11/15/05, Deal, Richard <rdeal at tigr.org> wrote:

...

I have a version of bb-xsnmp.pl running directly under hobbit. Most of the graphs already work with the hobbit server. I ripped out all the bb-combo.sh stuff as hobbit doesn't have combo.sh but that is just to get it to run under the hobbit server. I also added support for querying the hobbit server to see if a conn test is red to prevent hangs. This was a bigger deal as I would find that bb-xsnmp.pl would hang if a small non-important device failed such as a wireless access point and all the other devices would go purple instantly. I could give the patches but my version of bb-xsnmp.pl is based on a beta version I received from someone else that fixed other issues, so I will include the whole thing.

I put this in hobbitlaunch.cfg: [xsnmp] ENVFILE /local/HOBBIT/hobbit/client/etc/hobbitclient.cfg NEEDS hobbitd CMD /local/HOBBIT/hobbit/server/ext/bb-xsnmp.pl LOGFILE $BBSERVERLOGS/bb-xsnmp.log INTERVAL 5m I have it testing several APC UPS units with environment monitors (temperature graph doesn't work but monitoring does)

5 netapp servers for : cpu disk fans power temperature uptime Cpu, disk, and inode graphs work

65xx IOS routers/switches for : cpu uptime memory power fans temperature bgp Cpu graph works but temperature is totally unusable

65xx CatOS switches for : uptime memory power fans temperature

Cisco 12xx wireless access points for: cpu memory Cpu graph works.

I hope this helps BTW, I also run these from the Hobbit server: aironet.pl bb-ospf.pl bb-xsnmp.pl bb-brocade.pl bb-liebert.pl

And these on various clients: bb-cpu2.sh bb-prtdiag.sh emc-clarrion.sh ntpstat.sh bb-msgs.sh bb-raid.sh sge.sh bb-ntp.sh bb-vxcheck.sh meta

...

-----Original Message----- From: Dan Vande More [mailto:bigdan at gmail.com] Sent: Monday, November 14, 2005 5:36 PM To: hobbit at hswn.dk Subject: [hobbit] hobbit + bb-xsnmp.pl

Hey has anyone got the configs to get bb-xsnmp.pl to work with hobbit rrd and graphing? Henrik I saw you offered to do it here:

http://www.hswn.dk/hobbiton/2005/04/msg00080.html

But I can't find much else regarding it.

Thanks

Dan

To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk

To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk

To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk

To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk