[ofa-general] Bonding fail over not working

[Pradeep Satyanarayana]

Or Gerlitz wrote:
> Pradeep Satyanarayana <pradeeps at linux.vnet.ibm.com
> <mailto:pradeeps at linux.vnet.ibm.com>> wrote:
> 
>     I downloaded a recent version of Roland's git tree and tried IPoIB
>     bonding. Fail over does not seem to be working at all. I have tried
>     OFED 1.3.2 on a Rhel5 derivative and that (fail over) worked as
>     expected.
> 
>     Is this a known issue? Given that OFED 1.4 will be in sync with main
>     line kernel, is this an issue to be addressed in OFED 1.4 too? Has
>     any one else tried this out recently? My impression is that all
>     bonding patches were already upstream.
> 
> 
> I just tried ipoib/bonding with mainline kernel 2.6.27 and it works
> fine, as expected, see below. Can you repeat the exact sequence and see
> if it works for you, or send the settings that break bonding/ipoib on
> your system? I didn't use network scripts but this should be the issue
> if you use the directives that come with the ib-bonding package.

I just retested and it is indeed working as expected. My earlier conclusions
were erroneous. Between my experiments some one must have stepped on the cable and 
when I checked the cabling in the lab the port was disconnected. No wonder, the 
fail over did not occur as expected.

Sorry about the false report!

Pradeep