[OpenAFS] 1.4.0 on Solaris 10 x86/amd64

chas williams - CONTRACTOR chas@cmf.nrl.navy.mil
Mon, 14 Nov 2005 17:15:26 -0500


In message <20051111214453.GA9339@ccali28.in2p3.fr>,Loic Tortay writes:
>  avc->linkData
>  cpos <null string>
>  volnamep f/afs

i guess i shouldnt be surprised.  its never simple.  anyway, looking
at one of your trace again, this is pretty strange:

	time 452.000000, pid 344: RPC GetVolumeByName for 536871133 (at 0x9bb8e6)
	time -1551.-626552, pid 344: Analyze RPC op -1 conn 0x889315c8 code 0x0 user 0x0
	time -1491.-359488, pid 344: Analyze RPC op 2 conn 0x88931a50 code 0x0 user 0x64
	time 462.205880, pid 344: ProcessFS vp 0x8c2ab000 old len (0x0, 0x800) new len (0x9bb8f0, 0x800)

ok, its pretty unlikely that new len grew to this size.  its almost like
Length_hi has random contents.

	time 378.564965, pid 344: RPC GetVolumeByName for 536871133 (at 0x9bb826)
	time -1551.-626552, pid 344: Analyze RPC op -1 conn 0x889315c8 code 0x0 user 0x0
	time 175.749413, pid 344: GetdCache vp 0x8c2ab000 dcache 0x8c6465a0 dcache low-version 0x66, vcache low-version 0x66
	time 175.749413, pid 344: GetdCache tlen 0x800 flags 0x1 abyte (0x0, 0x0) Position (0x9bba70, 0x0)

and of course, Position is now wrong.

so, just to be certain about some things, the 1.4.0 release is broken
right (even after clearing the cache)?  and specifically, what options do
you pass to afsd?