Enhancement Request - show errors on link

[GoogleCodeExporter]

As a traffic Generator - Ostinato is great - but one thing I believe it's 
missing is any indication of errors on a link.

Just a general indication of CRC errors or something of the sort (stuff you'd 
see under Linux with ethtool -s ) would be GREAT.  One more row on the Ostinato 
data display.

You'd also want to clear the errors along with "bytes transmitted ..etc" 
counters so you could easily see if your link is working well.

I'd like to see this on both Windows and Linux versions - but right now believe 
the Linux one is the best  (since Ostinato tends to pump more traffic through 
it).

Thanks for your consideration.

-Werner H.

Original issue reported on code.google.com by whoehenr...@gmail.com on 2 Dec 2010 at 4:36

[GoogleCodeExporter]

Original comment by pstav...@gmail.com on 2 Dec 2010 at 5:26

• Changed state: Accepted
• Added labels: Type-Enhancement
• Removed labels: Type-Defect

[GoogleCodeExporter]

Srivats,

Hope all is well and HAPPY NEW YEAR to you.  Just curious - since it looks like 
you are "looking for things to add" - do you think getting a link error 
indication (Issue 28) might be in the cards for a future release?  I would 
think this shouldn't be too hard to implement (at least on Linux) since the 
info should be available.
Aside from this, and scripting support ;-) (to allow Ostinato to be a "canned" 
operation", something I can have anyone setup) this has most everything I need.

Much appreciate all your hard work and support.

Regards,
 Werner H. 

Original comment by whoehenr...@gmail.com on 11 Jan 2012 at 7:32

[GoogleCodeExporter]

@werner: Thanks for your wishes! Would the errors from the output of 'cat 
/proc/net/dev' work for you? Not sure if they are the same that 'ethtool -s' 
reports.

btw scripting support is not in the near future - I need to do some major 
refactoring and rearchitect parts of the code before I can start supporting 
scripting - but it's definitely in the plan.

Original comment by pstav...@gmail.com on 21 Jan 2012 at 8:47

[GoogleCodeExporter]

Srivats,

Sorry it took me a while to respond - I've been off at a conference. 

Yes - I think the error info from 'cat /proc/net/dev' would work fine.  The RX 
errs would be the primary one, but indicating RX drops and maybe fifo & frame 
would be good too.  Again - when you clear the Ostinato display, you want these 
statistics cleared also (at least on the Ostinato display).

I see that /proc/net/dev also has a column for TX errs - but don't know how 
that will ever show anything (I understand RX error, but not TX error?).

No worries on the scripting (we'd all love to have it, but realize that's a 
heavy lift).  Hopefully something like adding these errors isn't too big a 
change.

Regards,
 Werner

Original comment by whoehenr...@gmail.com on 1 Feb 2012 at 10:34

[GoogleCodeExporter]

@werner: I just pushed in a changeset to support rx error stats on Linux. Build 
from source and give it a shot. If you see any issues, let me know - I haven't 
had a chance to test it thoroughly

For now the Linux error stats are from /proc/net/dev (similar to the rx/tx 
stats), but in future this will be changed to use netlink and retrieve more 
detailed stats as well as link state.

Original comment by pstav...@gmail.com on 8 Feb 2012 at 4:51

[GoogleCodeExporter]

Srivats,

Great Job!!!  I just downloaded the updates and the Port Drops, errors ..etc 
now get displayed.  I disconnected my cable briefly and errors were nicely 
shown.  As requested, these get "zero'd" out when clearing the statistics - 
just like I wanted.

Well done!

Many thanks,
 Werner

Original comment by whoehenr...@gmail.com on 8 Feb 2012 at 11:57

[GoogleCodeExporter]

This issue was updated by revision 09dd87336e34.

Original comment by pstav...@gmail.com on 12 Feb 2012 at 7:26

[GoogleCodeExporter]

revision b781e68848cd added support for stats using netlink for Linux - link 
state info is now available

Original comment by pstav...@gmail.com on 10 Mar 2012 at 8:04

[GoogleCodeExporter]

Marking this issue as closed - any new error stats related enhancements should 
be raised as a new issue

Original comment by pstav...@gmail.com on 27 Mar 2014 at 3:31

• Changed state: Fixed