On Tue, 8 Nov 2005, Loic Tortay wrote:
According to chas williams - CONTRACTOR:
In message <[EMAIL PROTECTED]>,Loic Tortay writes:
According to
<https://lists.openafs.org/pipermail/openafs-info/2005-October/019683.html>,
Kris Kasner saw the problem on 1.3.81 too.
ok, try the 1.3.81 version then. i suspect its broken. looks like some
cleanup work went into the solaris client between .79 and .81. anyway
get .81 built and loaded on a 64 machine. before you try to access
afs, run 'fstrace setset'. try to access /afs. then run 'fstrace dump'
and send along the output. this should give us an idea where things
are going wrong.
I have been able to use the 1.3.81 release in 64 bit mode for very
basic testing:
[]
But the 1.3.82 release has the problem described in the ticket #22320.
#define AFS_64BIT_CLIENT 1
went from unconditionally defined to being defined only if __amd64 is
defined.
The thread-specific free packet queueing mech was added in .82 as well.
That's about it for stuff that touches the cache manager in any relevant
way.
Derrick
_______________________________________________
OpenAFS-info mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-info
