ok, great. I'm always finding stuff like that whenever someone comes up
with a new device. when to include it int the stats and when to skip it.
if you look at --netfilt you can explicitly tell collectl to ignore
various interfaces. that could help
-mark
On Wed, Oct 17, 2012 at 1:55 PM, Christopher Maestas <[email protected]>wrote:
> Thanks Mark. I'm doing well. Hope you are as well too!
>
> I found the issue. This is with an ovirt cluster (the free version of
> RedHat's virtualization engine) and there is a bridged interface called
> ovirtmgmt. I believe collectl was counting them twice (not that it was its
> fault). I was able to re-display only the main interface and am seeing
> (using netperf as the test case):
>
> # collectl -sN --netfilt p1p1
> waiting for 1 second sample...
>
> # NETWORK STATISTICS (/sec)
> #Num Name KBIn PktIn SizeIn MultI CmpI ErrsI KBOut PktOut SizeO
> CmpO ErrsO
> 7 p1p1: 12014 8169 1505 0 0 0 335 3890
> 88 0 0
> 7 p1p1: 12014 8225 1495 0 0 0 361 3943
> 93 0 0
> 7 p1p1: 12014 8186 1502 0 0 0 333 3886
> 87 0 0
>
> -cdm
>
>
>
>
> On Wed, Oct 17, 2012 at 11:04 AM, Mark Seger <[email protected]> wrote:
>
>> hey chris - great to hear from you and I hope things are going well for
>> you...
>>
>> as for the data types, they are indeed KBytes/sec and this IS saying you
>> are breaking the laws of physics so congratulations.
>> seriously though, I'm always saying collectl is pretty stupid and simply
>> reports what the kernel tells it to. therefore I'd be willing to bet
>> ethtool lying or perhaps more realistically the drive it lying to ethtool
>> about what speed it negotiated with the other end?!?
>>
>> the easiest way to verify all this is to look at /proc/dev/net before and
>> after the test OR do an ethtool -S before/after. Other than that I'm
>> afraid I can't offer much more.
>>
>> -mark
>>
>>
>> On Wed, Oct 17, 2012 at 12:18 PM, Christopher Maestas <
>> [email protected]> wrote:
>>
>>> Apologies for the formatting, but I have nodes on a 100Mb link and am
>>> trying to understand the actual reading here. If it's in Kbits, then it
>>> could make sense, but in KBytes, I don't see how 100Mb can do more than 12
>>> MB/s.
>>>
>>> #<----------Disks----------->
>>> <----------Network---------->
>>> #KBRead Reads KBWrit Writes KBIn PktIn KBOut PktOut
>>> 0 0 11448 90 21821 14098 575 6850
>>>
>>>
>>>
>>> This is on a FC17 system with the latest kernel update as of yesterday.
>>>
>>> $ rpm -q collectl
>>> collectl-3.6.3-2.noarch
>>>
>>> uname -r
>>> 3.6.1-1.fc17.x86_64
>>>
>>> $ ethtool p1p1
>>> Settings for p1p1:
>>> Supported ports: [ TP MII ]
>>> Supported link modes: 10baseT/Half 10baseT/Full
>>> 100baseT/Half 100baseT/Full
>>> 1000baseT/Half 1000baseT/Full
>>> Supported pause frame use: No
>>> Supports auto-negotiation: Yes
>>> Advertised link modes: 10baseT/Half 10baseT/Full
>>> 100baseT/Half 100baseT/Full
>>> 1000baseT/Half 1000baseT/Full
>>> Advertised pause frame use: Symmetric Receive-only
>>> Advertised auto-negotiation: Yes
>>> Link partner advertised link modes: 10baseT/Half 10baseT/Full
>>> 100baseT/Half 100baseT/Full
>>> Link partner advertised pause frame use: Symmetric
>>> Link partner advertised auto-negotiation: Yes
>>> Speed: 100Mb/s
>>> Duplex: Full
>>> Port: MII
>>> PHYAD: 0
>>> Transceiver: internal
>>> Auto-negotiation: on
>>> Cannot get wake-on-lan settings: Operation not permitted
>>> Current message level: 0x00000033 (51)
>>> drv probe ifdown ifup
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Everyone hates slow websites. So do we.
>>> Make your web apps faster with AppDynamics
>>> Download AppDynamics Lite for free today:
>>> http://p.sf.net/sfu/appdyn_sfd2d_oct
>>> _______________________________________________
>>> Collectl-interest mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/collectl-interest
>>>
>>>
>>
>
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
_______________________________________________
Collectl-interest mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/collectl-interest
