| | Posted 2/25/2008 7:19:29 AM | |
| Forum Newbie
       
Group: Forum Members Last Login: 3/20/2008 2:38:38 AM Posts: 3, Visits: 40 |
| | I have installed the v12 preview because we have problems to collect snmp-datas on cisco-gigabit switches with version 11.0.3 ==> this is the known 32-bit snmp wrap-around problem on high-capacity interfaces (like gigabit ....) On v12, the 64bit snmp counters should be supported. But the interface-performance monitors has still the same fault as Version 11.0.3. I have set the "poll interface traffic counter" to 64bit, but nothing changes. With the snmp-poller from Whatsup, I can't see the 64-bit HC-counter values. On the same server with the "iReasing-Mib browser" i can read the 64bit counters (s. also the attachements) Is the support for 64-bit counters not yet intergrated in the v12 preview? Thank's & Best Regards ervo |
| | | Posted 2/25/2008 8:30:19 AM | |
| 
Time Traveler
       
Group: Ipswitch Employees Last Login: Yesterday @ 3:07:15 PM Posts: 220, Visits: 335 |
| | ervo, The ability to poll using 64-bit counters has been added to the Interface Utilization Performance Monitor in v12. This feature is available in Preview 1. You simply need to ensure that an SNMPv2 or SNMPv3 credential has been assigned to the device and that the Interface Utilization Performance Monitor has been enabled for the device and is set to collect information using high capacity counters.
Have a good one, Christian Lawson QA Engineer - Network Monitoring Ipswitch, Inc. |
| | | Posted 2/29/2008 8:02:44 AM | |
| Forum Newbie
       
Group: Forum Members Last Login: 3/20/2008 2:38:38 AM Posts: 3, Visits: 40 |
| | ok, i have set the cisco snmp client to snmp version v2c. All settings on whatsup are now corrrect for 64bit HC-counters.You can see on the Wireshark capture-file that Whatsup receives all 5 minutes correct HC-counter values from cisco-device. But on the interface-utilization window the traffic-graphs goes still to 0 at this time?
|
| | | Posted 3/3/2008 4:47:40 PM | |
| 
Junior Member
       
Group: Ipswitch Employees Last Login: 8/18/2008 10:49:10 AM Posts: 19, Visits: 97 |
| | Hi Ervo, Thanks for the screen shots and packet capture. It looks like midway though the time period represented on that graph, you switched from 32-bit counters to high capacity. Is this true? What specifically on the graph is causing you concern? It appears to be graphing normally after you made the changes. Steve P |
| | | Posted 3/5/2008 1:58:16 AM | |
| Forum Newbie
       
Group: Forum Members Last Login: 3/20/2008 2:38:38 AM Posts: 3, Visits: 40 |
| | Hello Steve This gap occurs always on high traffic hours on all gigabit-interfaces. (We have had this problem also in version 11.0.3: Then we have opened a ticket (we have a service-agreement for this version) ==> In the answer was written that the perfomance-reports in version 11.0.3 doesn't support gigabit-interfaces). Now I have installed preview version 12. All settings are set to 64-bit-HC-counters. When the traffic on a gigabit-interface grows over about 12% load, the perfomance graph shows always this gap's. But the simultaneus trace's with wirehark in the background shows that whatsup collect the correct hc-counters? ervo
|
| | | Posted 3/5/2008 3:17:38 PM | |
| 
Junior Member
       
Group: Ipswitch Employees Last Login: 8/18/2008 10:49:10 AM Posts: 19, Visits: 97 |
| | Yes, there certainly seems to be a problem. May we get a copy of your database? If so, please send me an email or private message and we can get together on this. ~Steve P |
| | | Posted 3/14/2008 5:12:45 PM | |
| Supreme Being
       
Group: Ipswitch Employees Last Login: 8/26/2008 4:46:57 PM Posts: 87, Visits: 78 |
| Hello ervo, the issue was identical to the one in the thread titled "Still NO GIG interface support": http://forums.ipswitch.com/Topic41453-52-1.aspx
As I said in this thread, we found the problem and fixed it today. We were polling the right 64-bit counters but our bandwidth calculation formulas were invalid in that case when more than 4 billion bytes are going through the interface between two polls.
The fix will be in the next build that should be sent to the TPP in the next couple of weeks. I will definitively appreciate you trying it at that time. In the meantime, I would suggest to use the shortest possible polling interval you can.
Thank you again for finding that issue.
Sebastien Windal
Ipswitch, Inc |
| |
|
|