On May 21, 2015, at 4:14 AM, AMER <amer7...@hotmail.com> wrote: > Hello, > > Dear all > > when I tried to run the experiments uploaded on this link > https://github.com/clamy/cs244 > > I faced an error when I run them as the following: > error: integer out of range for 'L' format code > > I faced this error before with the Murphy Riplpox, however, when I installed > the latest updated version of RiplPox
My version *is* the latest version. :) > and taking the consideration of mininet and OvS versions it works fine. i.e., > the original riplpox. Unfortunately with the same setup, DCellController in > that uploaded experiments are showing that error again, I do not why the same > setup of mininet and OVS is working with original riplpox fine and it is not > working with DCell riplpox. Mininet and OVS are likely not the issue here. The likely issue (though it's hard to tell because you've posted an isolated error message and not a stack trace) is because old versions of RipLPOX (like the one included in that DCell project) are not compatible with later versions of POX. This particular issue is caused by RipLPOX using -1 as the buffer ID when there is no buffer to be specified. On more recent versions of POX, this is invalid, and one should use None instead. As I've mentioned before, you can either downgrade POX to a version that was current at the time of the version of RipLPOX that you're using, or you can fix the RipLPOX that you're using. For this specific bug, the problematic places should more or less show up in the stack trace, and it's also easy to find them by searching for -1 in the code. Lines 54 and 60 of the DCell riplpox.py are examples. You can also look at commits from my updated version of RipLPOX for guidance because I already fixed this a year ago: https://github.com/MurphyMc/riplpox/commit/c4d9272ac736aa61db2968927ea266454b1d6d56 -- Murphy > Best regards, > > > Eng Amer Alghadhban > COE > SANS-GCFW > CEH, SCNP, CCNA
