AIX OS Agent incorrect network stats

saffer

Hi,

I am running lpar2rrd-agent-6.10-0 on an AIX 6.1 server, and noticed, that the network stats are incorrect.
We have 5 networks as en3,en4,en5,en6,en7. 
By far the busiest network is en7, yet in the graphs, all of the networks report exactly the same utilisation.

Here is a text extract of the graph. [MB/sec] Int READ Avrg     Max     WRITE Avrg     Max     en5 9.13 15.37      34.87 133.66 en6 9.13 15.37      34.87 133.66 en7 9.13 15.37      34.87 133.66 en3 9.13 15.37      34.87 133.65 en4 9.13 15.37      34.87 133.65 Updated: Thu Dec 5 09:10:06 2019	[MB/sec] Int READ Avrg     Max     WRITE Avrg     Max     en3 9.06 14.98      33.82 72.22 en4 9.06 14.98      33.82 72.22 en5 9.06 14.98      33.82 72.22 en6 9.06 14.98      33.82 72.22 en7 9.06 14.98      33.82 72.22
Eth:pehdws09: last month [MB/sec] Int READ Avrg     Max     WRITE Avrg     Max     en3 9.23 14.91      33.43 47.92 en4 9.23 14.91      33.43 47.92 en5 9.23 14.91      33.43 47.92 en6 9.23 14.91      33.43 47.92 en7 9.23 14.91      33.43 47.92

Pavel

Hi,

I cannot see pictures, but send us this from the agent side:

ls -ltr /var/tmp/lpar2rrd*

tail /var/tmp/lpar2rrd*err

cat /var/tmp/lpar2rrd-agent1.out

https://upload.stor2rrd.com

Pavel

or email: support at stor2rrd dot com

saffer

Details uploaded

Mihail

Just to report the same issue. Details: The interfaces are VLAN, and each showing same traffic, which is actually total for the whole trunk port (LACP Link Aggregation).

Pavel

Hi,

just interfaces with IP assigned whould be presented. Has you vlan interfaces assigned an a IP?

do you seen any difference in entstat output?

entstat -d <interface>| grep "Bytes"

sleep 60

entstat -d <interface>| grep "Bytes"

same for the other interface on the other command line session

Is delta values more less same for each interface

Mihail

Hello,

Sorry , lost focus back in time,

To answer the questions:

- Yes, all vlan interfaces have IPs assigned,

- entstat - scripted (bash) in one session (I use single terminal):

#for i in {18..22}; do echo -n "en$i :" ;entstat -d en$i |grep "Bytes"; done; echo; sleep 60; for i in {18..22}; do echo -n "en$i :" ;entstat -d en$i |grep "Bytes"; done
en18 :Bytes: 8560464281236520                       Bytes: 15491069449339572
en19 :Bytes: 8560464283069830                       Bytes: 15491069450748900
en20 :Bytes: 8560464285235804                       Bytes: 15491069452216468
en21 :Bytes: 8560464287552680                       Bytes: 15491069453741884
en22 :Bytes: 8560464289454576                       Bytes: 15491069454957940

en18 :Bytes: 8560480570466930                       Bytes: 15491082397078928
en19 :Bytes: 8560480572907416                       Bytes: 15491082398891472
en20 :Bytes: 8560480575003638                       Bytes: 15491082400461136
en21 :Bytes: 8560480576646644                       Bytes: 15491082402098876
en22 :Bytes: 8560480578349896                       Bytes: 15491082403673920
#

Deltas are similar

Pavel

Hi,

yep, data is similar, we present just what we get from entstat, it is correct on our side as you can see it.

We have already got a fdew other users with the same issue.

It is always same, entstat reporting nearly same data throughput.

I do not know how it is possible, it is more question for AIX support.

Mihail

ok...

Pavel

let us know result if you raise a support case with IBM

Mihail

As a workaround i started using nmon as an additional source.