The newer OIDs come from a newer version of the MIB.
Incidentally, 1300 / 100 is 13. If it was displaying 0.13, then there is some maths failure within the polling code (or discovery is storing the wrong settings in the database).
The older one only has snAgentCpuUtilValue :
snAgentCpuUtilValue OBJECT-TYPE SYNTAX Gauge ACCESS read-only STATUS mandatory DESCRIPTION "The statistical CPU utilization in units of one-hundredth of a percent." ::= { snAgentCpuUtilEntry 4 }
adam.
On 2012-04-30 13:48, Matthias Cramer wrote:
Hi All
Sorry, if I did a "quick fix" but when I use it like this, the values correspond with what I see when i do a "sh cpu-utilization" or "sh cpu-utilization lp".
when I do a get in the main CPU it looks like that
orpheus:/opt/observium# snmpget -v2c -c community zelja.zh.as8758.net .1.3.6.1.4.1.1991.1.1.2.11.1.1.6.1.1.300 SNMPv2-SMI::enterprises.1991.1.1.2.11.1.1.6.1.1.300 = Gauge32: 1300
And that is relay 13% and not 0.13% as it showed for me before...
Something is weird here. So I looked at the source and the Database. Observium seams to use ".1.3.6.1.4.1.1991.1.1.2.11.1.1.4.1.1.300" and this returns whole precents. The SNMP Walk to snAgentCpuUtilEntry does not return any of the snAgentCpuUtil100thPercent oid's. I run a very recent version of Ironware, so either there is an other tree to get the "no depreciated" values or the MIB lies.
Anyways, my patch is wrong. And the MIB is probably too, at least snAgentCpuUtilValue does not return "one-hundredth of a percent".
In the meantime I also checked this against a RX8 and RX16 with the same result.
In which MIB have you found the snAgentCpuUtil100thPercent? It's not int any MIB provided with observium nor in any I have downloaded from Brocade.
Also in my MIB File STATUS is mandatory and not deprecated!
Regards
Matthias
On 30/04/12 13:20, Adam Armstrong wrote:
Both of the OIDs we use are in hundredths of a percent. There is the non-hundredths OID, but we don't use it.
snAgentCpuUtilValue OBJECT-TYPE SYNTAX Gauge32 MAX-ACCESS read-only STATUS deprecated DESCRIPTION "The statistical CPU utilization in units of one-hundredth of a percent. This value is deprecated. Users are recommended to use snAgentCpuUtilPercent or snAgentCpuUtil100thPercent instead." ::= { snAgentCpuUtilEntry 4 }
snAgentCpuUtilPercent OBJECT-TYPE SYNTAX Gauge32 MAX-ACCESS read-only STATUS current DESCRIPTION "The statistical CPU utilization in units of a percent." ::= { snAgentCpuUtilEntry 5 }
snAgentCpuUtil100thPercent OBJECT-TYPE SYNTAX Gauge32 MAX-ACCESS read-only STATUS current DESCRIPTION "The statistical CPU utilization in units of one-hundredth of a percent." ::= { snAgentCpuUtilEntry 6 }
Please be more careful when "fixing" things, and find out what the data actually should be!
adam.
On 2012-04-30 12:11, Tom Laermans wrote:
Matthias,
That sounds odd - the OID describes it as 100th of a Percent, and the patch was tested against Brocades before commit.
Maybe it's a bug?
Tom
On 30/04/2012 12:54, Matthias Cramer wrote:
Hi
The scale of the discovered CPU's on Ironware based geer is off by factor 100. This patch will fix that. Tested on Brocade XMR4000
Regards
Matthias
Index: includes/discovery/processors/ironware.inc.php
--- includes/discovery/processors/ironware.inc.php (revision 3098) +++ includes/discovery/processors/ironware.inc.php (working copy) @@ -15,11 +15,11 @@ { $usage_oid = ".1.3.6.1.4.1.1991.1.1.2.11.1.1.6." . $index; $usage = $entry['snAgentCpuUtil100thPercent'];
$precision = 100;
$precision = 1; } elseif ($entry['snAgentCpuUtilValue']) { $usage_oid = ".1.3.6.1.4.1.1991.1.1.2.11.1.1.4." . $index; $usage = $entry['snAgentCpuUtilValue'];
$precision = 100;
$precision = 1; } list($slot, $instance, $interval) = explode(".", $index);
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium