Nagios Plugins build under SPARC Solaris8 with SUNWspro CC fails

[Ralph.Grothe at itdz-berlin.de]

Hi Rob, Marc, Frank et alii,

I had to leave office yesterday earlier why I couldn't continue.

Thus only now my feedback.

Many thanks for your kind support.

Because I was tired of my fruitless efforts yesterday I think I
got a bit sloppy,
as can be seen from creating a Nagios plugin directory with mode
bits 0644 (where I meant 0755).
Another such flaw was giving configure --disable-openssl where it
should have been --without-openssl.
So I guess it was simply ignored.
But wouldn't you agree that it is rather inconsistent that
configure from the nrpe tarball on the other hand looks for the
--{enable|disable}-ssl switch.
Very confusing.
Anyway, I don't really require TLS/SSL for the Solaris managed
nodes.


This morning I quickly configured inetd on the Solaris box to run
nrpe,
and I was amazed that check* requests from my AIX Nagios server
were replied instantly.

$ uname -a
AIX daisy 3 4 0045CB7A4C00

$ libexec/check_nrpe -H zecke -c check_total_procs
PROCS OK: 119 processes


Of course I only tested the preconfigured check* commands in
nrpe.cfg right now,
and it still requires some fine tuning.
But I'm confident that at least our Solaris nodes can be
monitored by Nagios.
So I still have to fix the unwieldy HPUX nrpe daemon.

On the Solaris boxes we run Informix instances on Veritas Cluster
Server (VCS).

Although I know that it is very easy to check the status of
Informix DBs, and the VCS's out-of-the box Informix monitors are
ashamingly simple, I wonder if someone already has come up with a
check_informix or check_vcs/check_vxvm to save me the time coming
up with an own custom Perl script?

As you may have noticed I set up nrpe daemon to run as informix
rather than a less powerful nagios user.
I would rather prefer the latter for security concerns.
I only chose informix to be able to check the status of Informix
instances (viz. system() or exec*() of onstat,
or through Perl DBI::Informix).
How could that be better addressed?
Is there some suexec mechanism or would one set up sudo (or RBAC
in the case of Solaris)?
Suggestions welcome.


Frank,

thanks for the hints how to overcome the compiling catch with
check_icmp.
I will try your suggestion, and report.

Apropos, I so far haven't really understood what is expected as
PING_COMMAND string to render a working check_ping binary?
You have to admit that the documentation either in the Nagios
docs, or in configure's help screen is more than sparse.
I think the developers should have provided a terse example to
make things clear to dummies like me.

For instance on my Nagios Manager on AIX I defined this (because
I had fping already installed, from my Mon setup,
and I thought fping would perform better).

[nagios at daisy:~]
$ grep ^\#define\ PING_COMMAND
/opt/sw/nagios-plugins-1.4/config.h
#define PING_COMMAND "/usr/local/sbin/fping -q -C 3 -t 2000"

[nagios at daisy:~]
$ ls -l /usr/local/sbin/fping
-rwsr-xr-x   1 root     staff     140575 Apr 26 2004
/usr/local/sbin/fping


But this config rendered a broken check_ping binary.

On the other hand on the Solaris mananged node I simply chose the
path to Solaris' vanilla ping
(which, as you may know, without any options simply reports
"<node_name> is alive" to the astonishment of non-Solaris unix
admins).


# grep PING_COMMAND /usr/local/src/nagios-plugins-1.4/config.h
#define PING_COMMAND "/usr/sbin/ping"


But this doesn't work either 

[root at zecke:/usr/local/nagios/etc]
# ../libexec/check_ping -H themis -w 1000,50% -c 2000,80%
/usr/sbin/ping
CRITICAL - Could not interpret output from ping command


However, as said I don't really require the check_ping to be
executable from the managed nodes
but would be pleased to learn how its config was meant to be.

Regards
Ralph



> -----Original Message-----
> From: nagios-users-admin at lists.sourceforge.net
> [mailto:nagios-users-admin at lists.sourceforge.net]On Behalf Of
> Ralph.Grothe at itdz-berlin.de
> Sent: Tuesday, July 19, 2005 2:54 PM
> To: robmossrm at aol.com
> Cc: nagiosplug-help-admin at lists.sourceforge.net;
> nagios-users at lists.sourceforge.net
> Subject: RE: [Nagios-users] Nagios Plugins build under SPARC
Solaris8
> with SUNWspro CC fails
> 
> 
> Hi Rob,
> 
> I noticed that accidentally the separately mounted filesystem
> /usr/local/src
> got full during the build.
> Luckilly it was on an SDS volume set up as softpartition so
that
> I could growfs it.
> 
> I reran make distclean and reconfigured, this time disabling
the
> openssl stuff altogether.
> So there should be no need for the blooming compiler to search
> for such header files.
> 
> 
> # CC=/opt/SUNWspro/bin/cc ./configure
--with-nagios-user=informix
> --with-nagios-group=informix --disable-openssl
> --with-ping-command=/usr/sbin/ping
> 
> 
> This time my "make all" got a bit further, up to check_icmp
> (ok, if I get check_ping this time working, opposed to my
broken
> check_ping build on HPUX,
> I can ignore that, besides this is not a Nagios server, so
there
> should be little need for this node
> to ping anything).
> 
> 
> Here's the tail where cc aborted
> 
> 
> /opt/SUNWspro/bin/cc  -g  -L. -o check_procs  check_procs.o
> utils.o ../lib/libnagiosplug.a ../lib/li
> bcoreutils.a popen.o ../intl/libintl.a  -lgen -lsocket
> zecke --> Job output
> /opt/SUNWspro/bin/cc  -g  -L. -o check_icmp  check_icmp.o
> ../intl/libintl.a  -lgen -lsocket
> ild: (undefined symbol) gethostbyname -- referenced in the text
> segment of check_icmp.o
> ild: (undefined symbol) inet_addr -- referenced in the text
> segment of check_icmp.o
> ild: (undefined symbol) inet_ntoa -- referenced in the text
> segment of check_icmp.o
> *** Error code 5
> dmake: Fatal error: Command failed for target `check_icmp'
> Current working directory
> /usr/local/src/nagios-plugins-1.4/plugins
> Waiting for 1 job to finish
> zecke --> Job output
> /opt/SUNWspro/bin/cc  -g  -L. -o check_procs  check_procs.o
> utils.o ../lib/libnagiosplug.a ../lib/li
> bcoreutils.a popen.o ../intl/libintl.a  -lgen -lsocket
> *** Error code 1
> dmake: Fatal error: Command failed for target `all-recursive'
> Current working directory /usr/local/src/nagios-plugins-1.4
> *** Error code 1
> dmake: Fatal error: Command failed for target `all'
> [root at zecke:/usr/local/src/nagios-plugins-1.4]
> # df -k .
> Filesystem            kbytes    used   avail capacity  Mounted
on
> /dev/md/dsk/d103      342255  254652   59514    82%
> /usr/local/src
> 
> 
> Ok, I got enough check_* binaries compiled (the important ones
> for me being check_disk, check_procs, check_load)
> 
> # ls -F plugins|grep \*\$
> check_dhcp*
> check_disk*
> check_dummy*
> check_http*
> check_load*
> check_mrtg*
> check_mrtgtraf*
> check_nwstat*
> check_overcr*
> check_ping*
> check_procs*
> check_real*
> check_smtp*
> check_ssh*
> check_tcp*
> check_time*
> check_udp*
> check_ups*
> check_users*
> negate*
> urlize*
> 
> 
> Therefore I think I can get along with simply copying them
> manually into place
> 
> # mkdir -m 0644 -p /usr/local/nagios/libexec
> 
> # (cd plugins && cp $(ls -F|grep \*\$|tr -d \*)
> /usr/local/nagios/libexec/)
> 
> 
> Let's hope they are not broken.
> 
> Now I need to build the nrpe daemon.
> 
> Since I'm almost dead sure that this will start yet another
> debugging frenzy
> I'll be back soon to ask for further help...
 


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null