Weeding through the stream in wireshark doesn't sound that bad. You'd want to generate 1000pps to get 1ms accuracy though

------ Original Message ------
From: "Chuck McCown" <[email protected]>
To: [email protected]
Sent: 1/29/2018 10:20:38 AM
Subject: Re: [AFMUG] Failover / Recovery Time Testing?

Ping something that gives you <1mS pings then fail the circuit over and see what the ping time is on the next ping.

From:Christopher Gray
Sent: Monday, January 29, 2018 8:06 AM
To:[email protected]
Subject: Re: [AFMUG] Failover / Recovery Time Testing?

I want to use a tool that will time the "outage" automatically.

For example, switches with ERPS capability usually claim things like "sub 50ms recovery". How can I time that recovery on my actual system?

I picture running some continuous data stream between two points, and measuring any gaps in received data... but automatically instead of weeding through the stream in wireshark.







--------------------------------------------------------------------------------

On Mon, Jan 29, 2018 at 9:46 AM, Faisal Imtiaz <[email protected]> wrote:
Not sure what exactly are you asking about ?

How to break stuff ?
and or How to know when it is fixes ?

On a serious note, it depends on what you are testing..
e.g. one can easily simulate OSPF failover by turning a link / ethernet / radio port off.. and watching a traceroute (MTR or Pathping, Pingplotter etc etc )
and or watching the Router Logs

Regards.

Faisal Imtiaz
Snappy Internet & Telecom
http://www.snappytelecom.net

Tel: 305 663 5518 x 232 <tel:(305)%20663-5518>

Help-desk: (305)663-5518 <tel:(305)%20663-5518> Option 2 or Email: [email protected]

--------------------------------------------------------------------------------
From: "Christopher Gray" <[email protected]>
To: [email protected]
Sent: Monday, January 29, 2018 9:17:26 AM
Subject: [AFMUG] Failover / Recovery Time Testing?
What tools do you suggest for testing failover time / recovery time?
I've got several different failover mechanisms in use, and I'd like to start empirically testing them to have a better sense of their performance.

Thank you - Chris

Reply via email to