pcap replay error

[Felix Erlacher]

Dear trex folks,

I have been playing around with trex and encountered a strange issue.
To reproduce the issue I used the provided Fedora21 virtualbox image and
updated trex to the newest version (v2.39).

If I use the unchanged example http_simple.yaml in the command
t-rex-64 -f cap2/http_simple.yaml -d 1
and than look at the outcome via tcpdump -i enp0s8 everything looks fine.

If I change the pcap file in http_simple.yaml to the one attached
(httpSample.pcap) and use the same command again, the http response is
never seen by tcpdump (see tcpdump.pcap).
Also the trex packet counter shows less replayed packets than I would
expect.
The result is equivalent if I use the bp-sim64-debug command.

Is there something wrong with the pcap or am I just missing something?

thanks and greets
--
Felix

[Yaroslav Brustinov]

Hi,

The packet of response is too large (~11KB)

You have 2 options:

1. Fragment the packet in some way and save as new pcap
   
   
2. Use ASTF, which can fragment the data implicitly.
   Take a look at astf/nginx* files.

Thanks,

Yaroslav.

--
Felix

--
You received this message because you are subscribed to the Google Groups "TRex Traffic Generator" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trex-tgn+unsubscribe@googlegroups.com.
To post to this group, send email to trex...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/trex-tgn/b14e4739-26b6-1f6b-6aa8-443044c5f4db%40ccs-labs.org.
For more options, visit https://groups.google.com/d/optout.

[Felix Erlacher]

Thanks Yaroslav for the quick and helpful answer.
This solved my issue.
Just wondering, what is the packet size that trex is capable of sending?
It seems to be the standard 1514B, and it does not seem to be depended
on the interface MTU, I just tried that...

greets

felix

On 28/03/18 11:41, Yaroslav Brustinov wrote:
> Hi,
>
> The packet of response is too large (~11KB)
> You have 2 options:
>

> 1. Fragment the packet in some way and save as new pcap
>
> 2. Use ASTF, which can fragment the data implicitly.

> send an email to trex-tgn+u...@googlegroups.com
> <mailto:trex-tgn%2Bunsu...@googlegroups.com>.

> To post to this group, send email to trex...@googlegroups.com

> <mailto:trex...@googlegroups.com>.

> To view this discussion on the web visit
> https://groups.google.com/d/msgid/trex-tgn/b14e4739-26b6-1f6b-6aa8-443044c5f4db%40ccs-labs.org

> <https://groups.google.com/d/msgid/trex-tgn/b14e4739-26b6-1f6b-6aa8-443044c5f4db%40ccs-labs.org>.

> For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.
>
>

--
Felix Erlacher

ccs-labs.org/~erlacher
Key-ID:4EAC0959

[Yaroslav Brustinov]

cap2/imix_9k.yaml uses stl/ipv4_udp_9k.pcap, which has packet size 9230 (without FCS)

>     send an email to trex-tgn+unsubscribe@googlegroups.com

>     <mailto:trex-tgn%2Bunsu...@googlegroups.com>.
>     To post to this group, send email to trex...@googlegroups.com

>     <mailto:trex-tgn@googlegroups.com>.