Hi,
I'm trying to track the layer two activity of a group of ethernet
attached NT servers using snmpCollect to single collision frame
counter, output unicast and non-unicast packets and have combined these
values in the following mib expression which should track collisions to
output frames.
.1.3.6.1.2.1.10.7.2.1.4. .1.3.6.1.2.1.2.2.1.17. .1.3.6.1.2.1.2.2.1.18.
+ / 100 *
where :
collision frame counter = .1.3.6.1.2.1.10.7.2.1.4.
output unicast= .1.3.6.1.2.1.2.2.1.17.
output non-unicast=.1.3.6.1.2.1.2.2.1.18.
This expression forms a single line entry in mibExpr.conf as detailed
below:
Eth_Coll "blah" .1.3.6.1.2.1.10.7.2.1.4. .1.3.6.1.2.1.2.2.1.17.
.1.3.6.1.2.1.2.2.1.18. + / 100 *
When queried, the agent running on the NT server returns instances of
oid .1.3.6.1.2.1.10.7.2.1.4. as integers rather than the counter values
defined in the relevant RFC. I therefore used mib.coerce to modify the
default returned type as per the following extract from the file:
# dot3StatsSingleCollisionFrames
.1.3.6.1.2.1.10.7.2.1.4 COUNTER
When I run a data collection against the expression, the values
produced suggest that the oid values for .1.3.6.1.2.1.10.7.2.1.4 are
still being treated as an integer rather than the rate of change
values which would expected for a counter. As the following extract
from the collected data illustrates:
10s Mon Mar 15 11:03:14 bts-radb-s01 6417629.53
10s Mon Mar 15 11:03:24 bts-radb-s01 10077369.61
10s Mon Mar 15 11:03:34 bts-radb-s01 10326591.06
However, when I collect the single collision frame oid in isolation it
is correctly identified as a counter and I get the following output:
10s Mon Mar 15 11:09:14 bts-radb-s01 0.00
10s Mon Mar 15 11:09:24 bts-radb-s01 17.10
Has anyone had experience with this type of behaviour or have I
mis-configured a component of the collection process ?
I can't seem to identify any documentation that specifies the
relationship between mib.coerce and mib expressions, are there any
other references not covered in the dynatext books ?
Thanks in advance
Martin
|