Ok that helps a lot that they are the deltas. It's nice to have the ability to do this from the CLI. I like NS too but I don't like flipping back and forth between NS and CLI when I am looking into something so I am trying to figure out how to do these things both ways.
Have you worked much with OneView? In order to get this kind of info in OneView I have to enable "Interface statistics collection". I am nervous about doing this as I don't know its impact on the switches and the NS server. What's your take? On Wed, Jan 16, 2013 at 11:28 AM, Ernie Eaton Jr <[email protected]>wrote: > The port I chose is very lightly utilized (less than 1%). > > > > % utilization is the average utilization for the sample period. In my > case I showed an excerpt of 30 second intervals. The octets are bytes > for this interval. Same for packets and errors. These are the “deltas” not > the raw port counts so no need to do math. > > > > I find it really helpful to see the data in columns. To do this you must > disable the pager “set length 0” and then use the “wide” keyword with the > rmon command. > > > > Here is output from another switch which experienced “a burst” of activity > that brought utilization on the port up to 3 %. Stretch your email > window wide enough to see the table below properly formatted. Note how > easy it is to see the burst of activity in the octet or pkts column. The > Utilization column reflects this activity. We hold a about 25 minutes of > data for the 30 second intervals and 25 hours of data for the 30 minute > intervals. I find this much more useful than a single current utilization > counter because it allows me to see how traffic normally flows thru the > port and compare it to what is happening now (or just happened an hour > ago). > > > > > > > > NHSAL-IDF2-SW2(su)->show rmon history tg.11.1 wide > > > > Interface Index Interval Start Int Util > Octets Packets Broadcast Multicast Drop Evts Collisions > Jabbers CRC Errors Undersize Oversize Fragments > > ---------- ----- ------------------- ---- ---- -------------------- > -------------------- ---------- ---------- ---------- ---------- ---------- > ---------- ---------- ---------- ---------- > > tg.11.1 12201 2013-01-16 15:53:15 30 1% > 546989396 596569 2245 879 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:53:45 30 2% > 1117736383 1009168 1682 708 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:54:15 30 3% > 1845479905 1490266 2073 704 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:54:45 30 2% > 1298025010 1149088 1784 648 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:55:15 30 0% > 145667791 298144 1547 645 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:55:45 30 0% > 216321943 282335 1954 825 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:56:15 30 0% > 97992486 213491 1943 689 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:56:45 30 0% > 112225560 204147 1739 659 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:57:16 30 0% > 152944788 236587 1991 803 0 > 0 0 0 0 0 0 > > tg.11.1 12201 2013-01-16 15:57:45 30 0% > 312351337 373758 2252 734 0 > 0 0 0 0 0 0 > > > > > > *From:* John Kaftan [mailto:[email protected]] > *Sent:* Wednesday, January 16, 2013 11:10 AM > > *To:* Enterasys Customer Mailing List > *Subject:* Re: [enterasys] Mib for 1 min and 5 min average for port > throughput > > > > Ernie: > > > > Trying to make sense of this output below. Would I have to do the math to > get my number. i.e. subtract Octets and devide by seconds, or is there a > way to just see the average per interval from the CLI? I see % utilization > which could be helpful but it is 0 for both your example and mine. Perhaps > that doesn't show anything unless we get to a certain level of throughput. > I see how to do it now in NS and OneView. It would be nice to be able to > do it from the CLI too but if I have to do the math I'll stick to NS. > Thanks > > > > > > 1188391-800485 = > > Sample 1066 Interval Start: 23 days 5 hours 0 minutes 30 seconds > > Drop Events = 0 Undersize Pkts = 0 > > Octets = 800485 Oversize Pkts = 0 > > Packets = 5432 Fragments = 0 > > Broadcast Pkts = 3 Jabbers = 0 > > Multicast Pkts = 44 Collisions = 0 > > CRC Align Errors = 0 Utilization(%) = 0 > > > > Sample 1067 Interval Start: 23 days 5 hours 31 minutes 52 seconds > > Drop Events = 0 Undersize Pkts = 0 > > Octets = 1188391 Oversize Pkts = 0 > > Packets = 8913 Fragments = 0 > > Broadcast Pkts = 5 Jabbers = 0 > > > --More-- <space> next page, <cr> one line, <q> quit > > Multicast Pkts = 51 Collisions = 0 > > CRC Align Errors = 0 Utilization(%) = 0 > > > > > > > > JFK-D2(su)-> > > > > > > > > On Wed, Jan 16, 2013 at 10:09 AM, Ernie Eaton Jr <[email protected]> > wrote: > > By default (S/K/N series, not sure about the rest) we gather RMON history > every 30 seconds and every 30 minutes. We keep about 50 samples of each > for every port and make this available via CLI or SNMP. We’ve had some > problems calculating the bandwidth number in the past depending on the > ports operating speed. More recent code should function as described below: > > > > Our RMON implementation gathers data for both the receive and transmit > activity as if we had inserted a probe in the middle of the full duplex > cable. Future code will allow you to configure this behavior to monitor rx, > tx, or rx&tx. Today, since the data reported is rx&tx we calculate > bandwidth based on 2x the port speed. > > > > The “show rmon history” CLI offers the option to display the most recent > 30 seconds, 5 minutes or 25 minutes. This output is a summary of the 30 > second counters for the specified period. > > > > > > S8-202(su)->show rmon history wide ge.1.1 > > > > Interface Index Interval Start Int Util > Octets Packets Broadcast Multicast Drop Evts Collisions > Jabbers CRC Errors Undersize Oversize Fragments > > ---------- ----- ------------------- ---- ---- -------------------- > -------------------- ---------- ---------- ---------- ---------- ---------- > ---------- ---------- ---------- ---------- > > ge.1.1 1001 2013-01-16 10:34:27 30 0% > 11560 135 35 95 0 > 0 0 0 0 0 0 > > ge.1.1 1001 2013-01-16 10:34:57 30 0% > 11823 140 40 99 0 > 0 0 0 0 0 0 > > … > > > > > > S8-202(su)->show rmon history interval 5min ge.1.1 wide > > Interface Index Util Octets Packets > Broadcast Multicast Drop Evts Collisions Jabbers CRC Errors > Undersize Oversize Fragments > > ---------- ----- ---- -------------------- -------------------- ---------- > ---------- ---------- ---------- ---------- ---------- ---------- > ---------- ---------- > > ge.1.1 1001 0% 153020 1579 > 176 1089 0 0 0 0 > 0 0 0 > > > > > > > > > > *From:* Sillsby, David [mailto:[email protected]] > *Sent:* Wednesday, January 16, 2013 8:48 AM > > > *To:* Enterasys Customer Mailing List > *Subject:* Re: [enterasys] Mib for 1 min and 5 min average for port > throughput > > > > Hi Guys, > > > > Have you looked at RMON history? The default is a 30 minute window but it > is configurable if you must have 1 and 5 minute intervals. > > The CLI command is "show rmon history <port#>" > > > > > > Regards, > Dave Sillsby > Escalation Support Engineer > Enterasys Networks > > 9 Northeastern Blvd > > Salem, NH 03079 > > "There is nothing more important than our customers." > > > > On Wed, Jan 16, 2013 at 8:24 AM, Wyland, Tony <[email protected]> wrote: > > I too have asked about this functionality in the CLI and Enterasys points > to NetSight BUT the times that this capability is the most important is > when communications from NetSight may not be possible due to yuckness on > the switch/network. On non-Enterasys switches, this has been very helpful > to locate a source port of “junk” (extreme traffic or errors) that was > dragging a switch down. You find the port, disable it and work on finding > the real source of the problem while most of the users are become happy > quicker on a now working switch stack. > > Just a plug that it would still be a very useful feature that is > implemented on many other brands. Don’t get me wrong, we love our > Enterasys switches and there are many things done the way we like. > > > > Tony Wyland > > Messiah College > > > > *From:* Markus Kaiser [mailto:[email protected]] > *Sent:* Wednesday, January 16, 2013 3:29 AM > *To:* Enterasys Customer Mailing List > *Subject:* Re: [enterasys] Mib for 1 min and 5 min average for port > throughput > > > > Hi, > > > > if you use NetSight console, mark a port in the normal port flexview or > interface summary flexview, right-click and select "Port Monitor". > > Once the window and port information is shown wait at least 30 seconds > (refresh interval) and you will see the same information like at Cisco's > "show interface". > > > > Hope this helps. > > > > Kind regards, > > > > Markus > > > > > ___________________________ > > > On 16.01.2013, at 04:49, John Kaftan <[email protected]> wrote: > > Just came over from Cisco and we used to be able to do a "Show Interface" > command and see the 5 min average utilization for a specific port. It was > nice to be able to see that. I guess Cisco was just calculating that for > us as it went along. > > > > Does Enterasys do this? Is there a Mib I can hit for a 5 min port > utilization average? I don't see it in the CLI so I am thinking not. I > did find the flexview for port utilization but it looks like it is > calculating that after you open the flexview. I would like to be able to > just pull that info up on the fly like when doing the "show system > utilization" command. I don't want to wait 5 min to get the 5 min average. > > > > What I am doing is creating an ultimate interface flexview, i.e., one view > for all of the port info I am typically looking for. > > Thanks > -- > > John Kaftan > > IT Infrastructure Manager > > Utica College > > > > - --To unsubscribe from enterasys, send email to [email protected] with > the body: unsubscribe enterasys [email protected] > > > - --To unsubscribe from enterasys, send email to [email protected] with > the body: unsubscribe enterasys [email protected] > > > - --To unsubscribe from enterasys, send email to [email protected] with > the body: unsubscribe enterasys [email protected] > > > > - --To unsubscribe from enterasys, send email to [email protected] with > the body: unsubscribe enterasys [email protected] > > > - --To unsubscribe from enterasys, send email to [email protected] with > the body: unsubscribe enterasys [email protected] > > > > > > -- > > John Kaftan > > IT Infrastructure Manager > > Utica College > > > > - --To unsubscribe from enterasys, send email to [email protected] with > the body: unsubscribe enterasys [email protected] > > > - --To unsubscribe from enterasys, send email to [email protected] with > the body: unsubscribe enterasys [email protected] > > -- John Kaftan IT Infrastructure Manager Utica College --- To unsubscribe from enterasys, send email to [email protected] with the body: unsubscribe enterasys [email protected]
