Xymon with Bacula 🔗 link

Xymon Users,

I was wondering if anybody had experience with using xymon to monitor a
bacula backup server.
I am running a dedicated xymon server, and the bacula backup server is
offsite for backup reasons. I have a xymon client only installation on the
bacula server so I can receive cpu, disk, and the other system monitors. I
came across this xymon extension script (
http://www.revpol.com/xymon_bacula_check_script ) and I installed in on the
client side ( bacula backup server ). I followed the instructions (using
method 2 bconsole access & modifying the script to fit my environment ) and
the test of the xymon user ability to use bconsole works fine.

I am not seeing a backup or bacula section on the xymon server webpage. My
guess is that I have to add it somehow to the hosts.cfg file on the server
side but this is my first custom script / xymon extension that I had to use
so I'm not sure what the syntax should look like for the hosts.cfg. Should
I specify the full script name like this ( 155.x.x.101 backup # ssh
xymon_bacula_check.sh )? Perhaps there is something else im missing server
side that needs to be done like put the custom script in  server/ext ?

Appreciate the help
Thanks,
*                                                            *
*Joseph De Nicolo*
*Systems & Network Admin**istrator*
*Center for Complex Network Research <http://www.barabasilab.com>*
*Northeastern University*

On 01/10/13 13:30, Joseph De Nicolo wrote:

Hi Joseph,

You can run that (or any external script) manually like so:

# ~xymon/client/bin/xymoncmd ~xymon/client/ext/xymon_bacula_check.sh

and the new column should automatically and immediately appear for your bacula
server on your xymon interface.


If that works, then I have a couple guesses as to what might be wrong:

Is it possible that you cut and pasted that older information on that page
into the  ~xymon/client/etc/clientlaunch.cfg file on the bacula server?

That page was written back when Xymon was in mid process of a complete system
name change from hobbit to xymon. For a while, many files and variables had
kept their "hobbit" name.   So any mention of hobbit in there should probably
now be "xymon" assuming your running a relatively recent version of Xymon.

Also, the instructions on that page do not mention it, but I am pretty sure
that you need to restart the xymon client on the bacula server for it to
reread the clientlaunch.cfg file so it can start running that script every 5 mins.

I'll verify and then I will update that page. ;)

Hope this helps.

--
Bill Arlofski
Reverse Polarity, LLC

On 01/10/13 15:06, Joseph De Nicolo wrote:

the xymoncmd did the trick, I have to make sure that I don't have to do that
manually every time though.

(bringing this response back to the list)

Running it manually like that will run the script once. After (default) of 20
minutes that test will turn purple indicating that your Xymon server has not
seen an update for that test in a while.

If your bacula test turns purple, you will need to re-investigate your
settings on the client.

There should be a xymon_bacula_check.log log and a clientlaunch.log on the
bacula server in ~xymon/client/logs

Check to see if anything obvious shows up in those logs.

--
Bill Arlofski
Reverse Polarity, LLC

the xymon_bacula_check.sh script will run automatically now. The problem
was like you said, using $HOBBITCLIENTHOME instead of $XYMONCLIENTHOME when
you have to add those lines to the clientlaunch.cfg. I didn't catch it the
first time. I'm seeing the test perfectly fine on my xymon server website
now, even on server reboots.

Now I want to edit the actual bacula_check.sh script to show me more info.
I run multiple types of jobs so I want it to display Full, incremental,
differential, ect...
I believe I need to edit this line when it runs the list jobs command and
greps the data:
data=$(echo "list jobs" | $BCMD | $GREP "^|" | $TAIL -n$NUMJOBS | $SED -e
's/|\ /|/g'  \
     | $AWK -F"|" "{print \$2 \"\t\" \$3 \"\t\" \$4 \"\t\" \$7 \"\t\" \$8
\"\t\" \$9}" )

I'm just confused about those numbers, but I believe they represent the
column number of the catalog database correct? so to add job type and level
I just add columns 5 and 6 using the same format as you have here in the
script. And then I have to add those columns to the webpage by editing the
JOBHEADER line. This would be the correct way to go about this correct?

On Thu, Jan 10, 2013 at 3:23 PM, Bill Arlofski <user-0b8af203a56e@xymon.invalid>wrote:

-- 
*                                                            *

*Joseph De Nicolo*
*Secure Database Administrator
• *Center for Complex Network Research <http://www.barabasilab.com>*
*Northeastern University*

On 01/10/13 15:43, Joseph De Nicolo wrote:

Ok, good to know. That was my first thought.  Going to have to make an update
to that page to identify that issue.

I had already added the "Level" column a long time ago, but never updated that
script on the website. I just uploaded a new one now that includes some minor
changes to the one that was there.

Yes, that will work.  If you d/l the new script, you will see that I added a
"Level" column to the JOBHEADER var and column #6 to the data var.

I have this script running on 8 Bacula servers and I never bothered adding the
Job type (column 5 I think) since 99.99% of the jobs I see are "backup" jobs
and didn't think it was important. :)

Adding it to your status page should take you 20 seconds now that you have it
working.

Good luck!

--
Bill Arlofski
Reverse Polarity, LLC