perfparse-users

[Perfparse-users] Re: PerfParse Query

[Ben C. <Be...@cl...>]

Swarup,

I believe you do not have the library 'libmysqlclient', or this is in 
the wrong directory.  Can you please try and locate this lib:

# find / -name "libmysqlclient*"

When you have the location, please edit the Makefile:

LDFLAGS= -L/usr/local/lib/mysql -lmysqlclient -lm

Adjust the '-L' Linkage path for that which is correct for your setup.

Please let me know if this solves your problem.

Regards, Ben.

PS, I will would like to copy this email to the mailing list 
'perfparse-users' so that other users can see these comments.




Swarup Kumar wrote:

> Sir, 
>      I am trying to install perfparse on redhat enterprise. I have installed 
> nagios 1.2 in /usr/local/nms directory. Before compiling the perfparse, i 
> made the necessary PATH changes in config.h,Makefile and perfparse.cfg.
> I am already using mysql and have created a database named nms with the 
> command supplied by u in README.txt.
> I have installed Mysql 4.0.18 and running fine.
> 
> Now when i run perfparse command to start it, gives me the following error:
> 
> 
> [root@cruxonline bin]# ./perfparse
> ./perfparse: error while loading shared libraries: libmysqlclient.so.10: 
> cannot open shared object file: No such file or directory
> [root@cruxonline bin]#
> 
> 
> Please help
> 
> Thanks
> 
> Swarup Kumar
> INDIA
> 

[Perfparse-users] Re: PerfParse Query

[Ben C. <Ben...@ro...>]

Swarup,

Sorry but you do need to compile with --with-file-perfdata.  There is no 
other method to get Nagios to produce the data.

The following is a check-sheet of other Nagios related problems:

1. You are using the latest Nagios, 1.2.

2. You are using the latest nagios plugins.  Only these produce 
performance data.  You can check this by running one, like 'check_disk'. 
  If you have performance data, this should reply:

DISK OK - free space: / 8089 MB (22%); /boot 35 MB (76%) | 
/=8089MB;33312;35163;0;37014 /boot=34MB;40;42;0;45

You can see the performance data beyond the '|'.

3. You have compiled nagios with the option to produce performance data. 
  Using the '--with-file-perfdata'.  Ie:

./configure --with-file-perfdata

4. Set the permissions on the servicelog file as:
	-rw-rw-r-- nagios nagios
ie:
$ chmod 664 serviceperf.log
$ chown nagios.nagios serviceperf.log

5. Check 'nagios.cfg', ensure you have these lines somewhere:

process_performance_data=1
xpdfile_host_perfdata_file=/usr/local/nagios/var/hostperf.log
xpdfile_host_perfdata_template=$TIMET$\t$HOSTNAME$\t$OUTPUT$\t$PERFDATA$
xpdfile_service_perfdata_file=/usr/local/nagios/var/serviceperf.log
xpdfile_service_perfdata_template=$TIMET$\t$HOSTNAME$\t$SERVICEDESC$\t$OUTPUT$\t$PERFDATA$

6. In services.cfg, ensure this line is in all services you want to 
monitor:  (Or the generic parent table.)

process_perf_data 1

7. In Nagios 'Process Info' page. At the bottom of the Process 
Information box, ensure you have:

"Performance Data Being Processed?    Yes"


Regard, Ben

[Perfparse-users] Re: PerfParse Query

[Ben C. <Be...@cl...>]

Swarup,

I am glad you have it working.

It will add hosts as and when it finds data for them.  The reasons it=20
has not added all hosts can only be that there is not yet any data for th=
em.

You will want to run 'perfparse' frequently.  I find that once every ten=20
minutes is good.  You will find a file 'perfparse.sh' which should have=20
been moved into nagios/bin.  Run this from the cron every ten minutes.

This will add lost of data.  The main reason for perfparse is to have a=20
permanent record of your network.  My database of about one months has=20
3.6 million records in the perfdata_service_bin file, and the whole=20
database is about 4.5 Gb in size.  My best advise is to allow lots of=20
room, or store data for just a few hosts, or use a long check frequency=20
within Nagios.  Also use Nagios to check on the size of this disk.  You=20
also may want to control the MySQL logs, they grow much faster than the=20
data they log.

At least disks are cheep at the moment.  A 120Gb disk capable of holding=20
two years of our data is less than $100. (=A375 here in UK.)

There is not yet any method of deleting from this database.  If you wish=20
to delete, then you will have to manually.  Using 'mysql', enter=20
something like:

DELETE FROM perfdata_service_bin WHERE ctime < NOW() - INTERVAL 30 DAYS;
DELETE FROM perfdata_service_raw WHERE ctime < NOW() - INTERVAL 30 DAYS;

If you don't use the 'raw' data, which is not used by any of PerfParse,=20
then simply enter:

DELETE FROM perfdata_service_raw;

Or, if you just want to delete data from a specific host, enter:

DELETE FROM perfdata_service_bin WHERE host =3D 'host_name' AND ctime <=20
NOW() - INTERVAL 30 DAYS;

You can add any of these commands into perfparse.sh.  Eg:

echo "DELETE FROM perfdata_service_raw;" | mysql -u user -ppass -D=20
database -h hostname

Deleting the MySQL logs is simple.  Find the data or log directory and=20
enter:

# rm database-bin.0*
# rm access.log

Where 'database' is the name of your host, which may be changed in my.cfg.

DO NOT EVER delete database-bin.index !!!!

These features will be added to the product at some time in the (near)=20
future.

Regards, Ben.

PS, I have posted this reply to perfparse-users mailing list for the=20
interest of other users.