[OpenAFS-devel] AIX6 support - any progress?

[Niklas Edmundsson]

On Fri, 12 Dec 2008, Derrick Brashear wrote:

>>> There is an AIX 6.1 binary on the 1.4.8 release page
>>>
>>> http://www.openafs.org/pages/release/openafs-1.4.8.html
>>>
>>
>> Oh. That's what I get for being out of the loop, I've even managed to miss
>> that release announcement.
>>
>> However, things doesn't seem perfect since afsd segfaults on me upon
>> startup. Debug symbols seems stripped too, so the coredump doesn't do much
>> good.
>
>
> Leaving debug symbols in release binaries is... debatable.
>
> It did work on the test system.

Weird. I recompiled from the 1.4.8 tarball with --enable-debug (which 
leaves optimization enabled, it only adds debug symbols and disables 
stripping the binaries) and now afsd doesn't segfault on me.

If I do the stupid thing and define the CellServDB without any server 
entries and start afsd without dynroot/afsdb-support the machine bombs 
in kernel space. I suspect that there is a trivial check missing 
somewhere in there ;)

Anyhow, I seem to have some basic working AFS-molecules in my AIX6 box 
so I'll distract myself with some setup and installation tuning for a 
while until I've worked off enough backlog to have time to do some 
more debugging.

Hints on where to look for bugs or medium sized ogres are appreciated 
though, although I won't have time now I'll get to my mail list 
backlog eventually ;)

/Nikke
-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
  Niklas Edmundsson, Admin @ {acc,hpc2n}.umu.se     |    nikke@hpc2n.umu.se
---------------------------------------------------------------------------
  Tagteam:  A bunch of people thinking up taglines.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=