Sflow failed ?

[harvestbc]

Hi support,

I'm testing out this product and monitoring a powerconnect switch using sflow.

There are data coming in, however, it seemed to crash and stop collecting data when i recieve this error (/var/log/message):

localhost kernel: sfcapd[39202] general protection ip:7faeae48a7fe sp:c69afbee5ac39e7b error:0 in libc-2.12.so[7faeae4580000+18b000]


Seems like the only way to get the collection of data going again, i've to click on the restart button for this source on the NAA website.

*also, the stop button doesn't seem to be responding for a sflow type source

Any pointers what could be causing this error?

[sreinhardt]

At the moment I do not know of any issues specifically with powerconnect devices or sflow in regards to crashing nna collectors. Unfortunately it seems that the crash was more in a libc library and not necessarily with the nna collector based on your output there.

You are correct, at the moment if a collector is killed or stopped in some way, NNA does not attempt to restart it, I think this would be a very useful feature request, and will submit an internal one for it.

I will have to check on the sflow stop not working, this really shouldn't have any difference if it's an sflow or netflow collector, but we can do some testing and see!

[harvestbc]

Thanks.

My test environment is as follows if you need the info:

Switch Type: Powerconnect 6248
Switch Configuration: Stacked
Polling 30secs with a sampling size of 1024

[sreinhardt]

So far I am unable to replicate the same issue with stopping the collectors. What version of NNA are you running and on what OS?

[harvestbc]

I'm running:

NNA latest version - 2014R1.9
OS - Centos 6.5 64bit

[sreinhardt]

OK thanks, I will setup that same system and see what happens!