| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 3593.1 | "Drops" might mean dropped packets.... | NETCAD::BATTERSBY | Don't use time/words carelessly | Wed Jun 05 1996 12:39 | 4 |
| "Drops" is probably short for "dropped packets", IE: packets which
were ignored or rejected.
Bob
|
| 3593.2 | | NETCAD::GALLAGHER | | Wed Jun 05 1996 17:45 | 24 |
| etherStatsDropEvents OBJECT-TYPE
SYNTAX Counter
ACCESS read-only
STATUS mandatory
DESCRIPTION
"The total number of events in which packets
were dropped by the probe due to lack of resources.
Note that this number is not necessarily the number of
packets dropped; it is just the number of times this
condition has been detected."
::= { etherStatsEntry 3 }
etherHistoryDropEvents OBJECT-TYPE
SYNTAX Counter
ACCESS read-only
STATUS mandatory
DESCRIPTION
"The total number of events in which packets
were dropped by the probe due to lack of resources
during this sampling interval. Note that this number
is not necessarily the number of packets dropped, it
is just the number of times this condition has been
detected."
::= { etherHistoryEntry 4 }
|
| 3593.3 | Thanks | KERNEL::CHOPRAA | | Thu Jun 06 1996 09:38 | 9 |
| Thanks a lot for the quick reply
I thought that it could mean a resource problem but then decided that the probe should never
really run out of resources when looking at a network with very low utilisation - 4.8 (as per
the probe) in some cases.
Does this indicate a problem somewhere or is this just the way the probe works?
Anil
|
| 3593.4 | Bursts? | NETCAD::GALLAGHER | | Thu Jun 06 1996 10:04 | 20 |
| >Does this indicate a problem somewhere or is this just the way the probe works?
It's just the way the probe works. The counter increments when the probe is
too busy. Some of the ways you can get the counter to increment are:
- Monitor a 'busy' network (I don't know how many packets per
second the probe can handle, so I can't tell you what 'busy'
really means.)
- Monitor a network with very bursty traffic.
- Try to monitor a lot of stuff (e.g. enable a lot of domains).
The more you try to monitor, the more work the probe has to do,
and the greater the chance of it dropping packets.
At 4.8% utilization, I'd guess that you might have bursty traffic - perhaps
broadcast or multicast storms. Do you see and utilization spikes? Do
the broadcast and multicast packet counters seem to spike?
-Shawn
|
| 3593.5 | It all makes sense now :-| | KERNEL::CHOPRAA | | Thu Jun 06 1996 12:12 | 14 |
| >> Do the broadcast and multicast packet counters seem to spike?
YUP. The reason we were looking at the network was to resolve an intermittent connectivity
problem. With the information you gave me in .4, it seems logical to assume that there is
some sort of problem with "utilisation spikes" on an other-wise quiet network.
This would explain the packets being dropped even thought the utilisation on the network is
very low.
It all makes sense now :-|
Thanks very much for the help.
Anil
|