SNMP - Get Processor Information does not work properly

uidzer0 · Post by **uidzer0** » Fri Sep 21, 2007 2:17 pm

I am trying to add in the data query to run against a RH WS3 server (server24 below) and I am not getting any results from it in terms of CPU.

OS: Redhat WS3
Cacti: 0.8.6j (fully patched)
php: 5.2.4
rrdtool: 1.2.23
net-snmp: 5.0.9

Data Query Debug Information

+ Running data query [9].
+ Found type = '6 '[script query].
+ Found data query XML file at '/path/to/cacti/resource/script_server/host_cpu.xml'
+ XML file parsed ok.
+ Executing script for list of indexes '/path/to/php/bin/php -q /path/to/cacti/scripts/ss_host_cpu.php server24 15 2:161:500:COMMUNITY:::MD5::[None] index'
+ Executing script query '/path/to/php/bin/php -q /path/to/cacti/scripts/ss_host_cpu.php server24 15 2:161:500:COMMUNITY:::MD5::[None] query index'
+ Found data query XML file at '/path/to/cacti/resource/script_server/host_cpu.xml'
+ Found data query XML file at '/path/to/cacti/resource/script_server/host_cpu.xml'
+ Found data query XML file at '/path/to/cacti/resource/script_server/host_cpu.xml'

host_cpu.xml calls ss_host_cpu.php and looking in there, I see the following:

$oids = array(
"index" => ".1.3.6.1.2.1.25.3.3.1",
"usage" => ".1.3.6.1.2.1.25.3.3.1"
);

Output from snmpwalk command based on those:

[root@cactiserver]# snmpwalk -v 2c -c COMMUNITY server24 .1.3.6.1.2.1.25.3.3.1
HOST-RESOURCES-MIB::hrProcessorFrwID.769 = OID: SNMPv2-SMI::zeroDotZero

UPDATE:

I decided to try a newer version of net-snmp (5.2.4) and I got the same output.

uidzer0 · Post by **uidzer0** » Mon Sep 24, 2007 10:02 am

I am still trying to figure out why I cannot get any information back from the SNMP - Get Processor Information. I have attached two pictures.. Any information or further troubleshooting steps would be extremely helpful.

kernelphr34k · Post by **kernelphr34k** » Mon Sep 24, 2007 10:48 am

I'm in the same boat. Can't get SNMP processor info for my localhost.

suse enterprise 10 linux(2.6.16 kernel)
cacti 0.8.6j + patches

net-snmp = 5.3.0.1-25.2
php5 = 5.1.2-29.5
rrdtool = 1.2.12-13.2

+ Running data query [9].
+ Found type = '6 '[script query].
+ Found data query XML file at '/srv/www/htdocs/cacti-0.8.6j/resource/script_server/host_cpu.xml'
+ XML file parsed ok.
+ Executing script for list of indexes '/usr/bin/php5 -q /srv/www/htdocs/cacti-0.8.6j/scripts/ss_host_cpu.php tcp:127.0.0.1 1 1:161:500:COMM_STRING:::MD5::[None] index'
+ Executing script query '/usr/bin/php5 -q /srv/www/htdocs/cacti-0.8.6j/scripts/ss_host_cpu.php tcp:127.0.0.1 1 1:161:500:COMM_STRING:::MD5::[None] query index'
+ Found data query XML file at '/srv/www/htdocs/cacti-0.8.6j/resource/script_server/host_cpu.xml'
+ Found data query XML file at '/srv/www/htdocs/cacti-0.8.6j/resource/script_server/host_cpu.xml'
+ Found data query XML file at '/srv/www/htdocs/cacti-0.8.6j/resource/script_server/host_cpu.xml'

uidzer0 · Post by **uidzer0** » Mon Sep 24, 2007 10:59 am

I just tried a downgrade of cacti, from 0.8.6j to 0.8.6h and still had no success. I'm considering downgrading to a 4.x version of ucd-snmp and see if that makes a difference.

uidzer0 · Post by **uidzer0** » Mon Sep 24, 2007 12:07 pm

Update:

Installed a 4.x version of ucd-snmp, changed the setting in cacti to use ucd-snmp 4.x and also changed the paths to the snmp* binaries to reflect the ucd-snmp installation. The 'Get processor Information' data query still does not work.

kernelphr34k · Post by **kernelphr34k** » Tue Sep 25, 2007 11:58 am

update for me.

I hope you don't mind me posting in this thread.. Sounds like we have almost the same issue...

I just added a windows box to devices, and was able to get the cpu info for the 2 procs on the box. Yet I'm unable to obtain the cpu info on the local linux box...

elgoretto · Post by **elgoretto** » Wed Sep 26, 2007 4:06 am

"Me too"

Many of linux SNMP devices (net-snmp versions 5.0.9, 5.2.3, or 5.4.1) don't show any item to be graphed (all Red Hat, Debian and Ubuntu boxes with same Data Query Debug Information as uidzer0, while Gentoo servers are OK).

re-Edit: During my last attempt to compile and install net-snmp 5.4.1 on a vmware ESX server host (OS like a RedHat distro), CPU shows up, so it' works. I'll try to compile it again but on a RedHat AS 4.

Btw, the "ucd/net - CPU Usage" works.
I use 0.8.6j,

--
edit: oops, I forgot: getting all CPUs on Windows machines works, as on Gentoo ones.

uidzer0 · Post by **uidzer0** » Thu Sep 27, 2007 7:57 am

elgoretto wrote:"Me too"
Many of linux SNMP devices (net-snmp versions 5.0.9, 5.2.3, or 5.4.1) don't show any item to be graphed (all Red Hat, Debian and Ubuntu boxes with same Data Query Debug Information as uidzer0, while Gentoo servers are OK).

re-Edit: During my last attempt to compile and install net-snmp 5.4.1 on a vmware ESX server host (OS like a RedHat distro), CPU shows up, so it' works. I'll try to compile it again but on a RedHat AS 4.

Btw, the "ucd/net - CPU Usage" works.
I use 0.8.6j,

--
edit: oops, I forgot: getting all CPUs on Windows machines works, as on Gentoo ones.

Well, this is absolutely driving me crazy. I can not get the data query to work at all and it's my understanding that it has to work in order to get the multi cpu graphs to work. I've tried reinstalls, building from source, installing via rpm, etc, etc.. (the servers are RH WS3 boxes)

elgoretto · Post by **elgoretto** » Thu Oct 18, 2007 10:06 am

Sorry, I recently hadn't time enough to go on this problem.

So, I installed a new RedHat 4.5 testbox. I have to upgrade the net-snmp package provided because diskio isn't enabled by default on all RH4.x (what a pity).
See my previous post for my former tests with net-snmp and various distros.
[edit: note that the provided 5.1.2 net-snmp is correct regarding the SNMP - Get Processor Information datasource. It works "as is".]

Well I finally managed to compile net-snmp 5.4.1 on a RH AS 4.5 and get it working. For this to work, I had to disable SElinux, which seems to block any SNMP answer from the RH box.
And hurrah!, the SNMP - Get Processor Information datasource returned something

Code: Select all

+ Running data query [9].
+ Found type = '6 '[script query].
+ Found data query XML file at '/var/www/localhost/htdocs/cacti/resource/script_server/host_cpu.xml'
+ XML file parsed ok.
+ Executing script for list of indexes '/usr/bin/php -q /var/www/localhost/htdocs/cacti/scripts/ss_host_cpu.php 10.231.226.100 12 2:161:500:LABTECH_RO:::MD5::[None] index'
+ Executing script query '/usr/bin/php -q /var/www/localhost/htdocs/cacti/scripts/ss_host_cpu.php 10.231.226.100 12 2:161:500:LABTECH_RO:::MD5::[None] query index'
+ Found item [hrProcessorFrwID='0'] index: 0

If this may help:

modify /etc/selinux/config and set SELINUX=disabled
not sure but reboot could be a good idea
uninstall previous net-snmp package
compile net-snmp with options like these ones (any suggestion is welcome):
./configure --prefix=/ --exec_prefix=/usr --with-mib-modules=ucd-snmp/diskio --disable-set-support --with-default-snmp-version=2 --with-sys-contact="root@Unknown" --with-sys-location="Unknown" --with-logfile="/var/log/snmpd.log" --with-persistent-directory="/var/net-snmp" --disable-mfd-rewrites --enable-as-needed --enable-shared --without-libwrap --without-rpm --without-bzip2 --without-zlib --without-elf --disable-manuals --disable-scripts
make install and enjoy

gangstafish · Post by **gangstafish** » Tue Oct 23, 2007 11:20 am

I am getting this same issue, so frustrating. I have a few machines with almost the same exact build except for versions on snmp. The machines with snmp v5.1.3 work fine, the versions of snmp v5.2+ don't work a lick.

On the machines that do give back cpu information you see the expected snmpwalk:
+ Found item [hrProcessorFrwID='0'] index: 0
+ Found item [hrProcessorFrwID='1'] index: 1

On the machines that don't work, you get no found items, but no errors either. It seems like versioning and I feel like I'am missing something extremely easy.

prajjwal · Post by **prajjwal** » Thu Oct 25, 2007 9:24 am

Hi,
I have been trying to get cacti to poll some computers I have access to here. Other machines work fine, but I am getting cpu graphing errors on a RHEL 4 machine. The computer is a Dual Core AMD Opteron(tm) Processor 880, and there are multiple CPUs. However, when I run the data query in debug mode, I get the following:

+ Running data query [9].
+ Found type = '6 '[script query].
+ Found data query XML file at '/var/www/html/cacti/resource/script_server/host_cpu.xml'
+ XML file parsed ok.
+ Executing script for list of indexes '/usr/bin/php -q /var/www/html/cacti/scripts/ss_host_cpu.php xxxx.xxxxx.xxx 4 2:161:500:xxxxxxx:::MD5::[None] index'
+ Executing script query '/usr/bin/php -q /var/www/html/cacti/scripts/ss_host_cpu.php xxxx.xxxxx.xxx 4 2:161:500:xxxxxxx:::MD5::[None] query index'
+ Found data query XML file at '/var/www/html/cacti/resource/script_server/host_cpu.xml'
+ Found data query XML file at '/var/www/html/cacti/resource/script_server/host_cpu.xml'
+ Found data query XML file at '/var/www/html/cacti/resource/script_server/host_cpu.xml'

When I run snmpwalk, I can get the following output:
[prajjwal@xxxxx ~]$ snmpwalk xxxxx -v2c -c $community .1.3.6.1.2.1.25.3.3.1
HOST-RESOURCES-MIB::hrProcessorFrwID.768 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorFrwID.769 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorFrwID.770 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorFrwID.771 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorFrwID.772 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorFrwID.773 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorFrwID.774 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorFrwID.775 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorLoad.768 = INTEGER: 2
HOST-RESOURCES-MIB::hrProcessorLoad.769 = INTEGER: 1
HOST-RESOURCES-MIB::hrProcessorLoad.770 = INTEGER: 37
HOST-RESOURCES-MIB::hrProcessorLoad.771 = INTEGER: 55
HOST-RESOURCES-MIB::hrProcessorLoad.772 = INTEGER: 1
HOST-RESOURCES-MIB::hrProcessorLoad.773 = INTEGER: 1
HOST-RESOURCES-MIB::hrProcessorLoad.774 = INTEGER: 7
HOST-RESOURCES-MIB::hrProcessorLoad.775 = INTEGER: 9

There are no problems with other graphs. Can someone point me in the direction?

Thanks,
Prajjwal

Post by **TheWitness** » Thu Oct 25, 2007 10:13 pm

It would appear to me that the Agent on your box is hosed. You need to look at both your snmpd.conf file and the possibly upgrade/downgrade your agent. Because it is truely FUBAR.

TheWitness

flashdumper · Post by **flashdumper** » Thu Nov 15, 2007 2:26 am

Hello. I've the similar prob. What if i resolve an INTEGER value, but I want that value in to represent on the Graph? How can I do that?

fmangeant · Post by **fmangeant** » Thu Nov 15, 2007 2:42 am

Hi

IMHO with Net-SNMP you should use the "ucd/net - CPU Usage" graph template instead of the "SNMP - Get Processor Information" data query.

To check if you'll get the proper values, try that :

Code: Select all

$ snmpwalk -v 1 -c yourcommunity xx.xx.xx.xx systemStats
UCD-SNMP-MIB::ssIndex.0 = INTEGER: 1
UCD-SNMP-MIB::ssErrorName.0 = STRING: systemStats
UCD-SNMP-MIB::ssSwapIn.0 = INTEGER: 0 kB
UCD-SNMP-MIB::ssSwapOut.0 = INTEGER: 0 kB
UCD-SNMP-MIB::ssIOSent.0 = INTEGER: 17 blocks/s
UCD-SNMP-MIB::ssIOReceive.0 = INTEGER: 0 blocks/s
UCD-SNMP-MIB::ssSysInterrupts.0 = INTEGER: 100 interrupts/s
UCD-SNMP-MIB::ssSysContext.0 = INTEGER: 37 switches/s
UCD-SNMP-MIB::ssCpuUser.0 = INTEGER: 1
UCD-SNMP-MIB::ssCpuSystem.0 = INTEGER: 0
UCD-SNMP-MIB::ssCpuIdle.0 = INTEGER: 97
UCD-SNMP-MIB::ssCpuRawUser.0 = Counter32: 72522780
UCD-SNMP-MIB::ssCpuRawNice.0 = Counter32: 6235299
UCD-SNMP-MIB::ssCpuRawSystem.0 = Counter32: 2867141
UCD-SNMP-MIB::ssCpuRawIdle.0 = Counter32: 2355462838
UCD-SNMP-MIB::ssCpuRawWait.0 = Counter32: 19555889
UCD-SNMP-MIB::ssCpuRawKernel.0 = Counter32: 0
UCD-SNMP-MIB::ssCpuRawInterrupt.0 = Counter32: 85304
UCD-SNMP-MIB::ssIORawSent.0 = Counter32: 955122286
UCD-SNMP-MIB::ssIORawReceived.0 = Counter32: 493880696
UCD-SNMP-MIB::ssRawInterrupts.0 = Counter32: 2719804467
UCD-SNMP-MIB::ssRawContexts.0 = Counter32: 601861088
UCD-SNMP-MIB::ssCpuRawSoftIRQ.0 = Counter32: 93380
UCD-SNMP-MIB::ssRawSwapIn.0 = Counter32: 8164
UCD-SNMP-MIB::ssRawSwapOut.0 = Counter32: 22437

elgoretto · Post by **elgoretto** » Thu Nov 15, 2007 11:21 am

Ok, I managed to build a net-snmp-5.4.1 rpm package, and successfully deployed it on a RHEL 4.5, with cacti being happy.

As already said, upgrading the snmp agent solved the problem discussed here.

SNMP - Get Processor Information does not work properly

SNMP - Get Processor Information does not work properly

Who is online