[OpenAFS] dkms module for openafs

Thanks to the efforts from Pascal, we now have a dkms package for openafs in mandrake contrib, making the compilation of libafs module far easier than before. Package: ftp://ftp.free.fr/pub/Distributions_Linux/Mandrakelinux/devel/cooker/i586/media/contrib/dkms-libafs-1.3.82-2mdk.i586.rpm Spec: h

Re: [OpenAFS] heavy troubles with openafs 1.3.79 and kernel 2.6.11

chas williams - CONTRACTOR wrote: In message <[EMAIL PROTECTED]>,Guillaume Rousse writes: EIP is at __inode_queue_event+0x18/0x60 eax: f33c1000 ebx: fff8 ecx: f33c1000 edx: esi: 0004 edi: 0020 ebp: e8d3bf8c esp: e8d3bf74 ds: 007b es: 007b ss: 0068 Proc

Re: [OpenAFS] heavy troubles with openafs 1.3.79 and kernel 2.6.11

Guillaume Rousse wrote: any oops in dmesg? Right, I forgot: Found system call table at 0xc037c8c0 (pattern scan) Starting AFS cache scan...old style would have needed 108 contiguous bytes found 2146 non-empty cache files (8%). Unable to handle kernel NULL pointer dereference at virtual

Re: [OpenAFS] heavy troubles with openafs 1.3.79 and kernel 2.6.11

Guillaume Rousse wrote: i noticed that a change was revered in src/rx/rx.c you might try getting the latest version from cvs instead. This one ? http://www.openafs.org/cgi-bin/wdelta/rollback-20050315 I'll try, thanks. Apparently not. Release 1.3.79 of openafs include revision 1.58.2.6, of

Re: [OpenAFS] heavy troubles with openafs 1.3.79 and kernel 2.6.11

chas williams - CONTRACTOR wrote: In message <[EMAIL PROTECTED]>,Guillaume Rousse writes: Any attempt to browser a afs filesystem segfault: [EMAIL PROTECTED] ~]$ ll /mnt/afs/inria.fr/rocq/home/atoll/common/cvs Segmentation fault any oops in dmesg? Right, I forgot: Found system call ta

[OpenAFS] heavy troubles with openafs 1.3.79 and kernel 2.6.11

Any attempt to browser a afs filesystem segfault: [EMAIL PROTECTED] ~]$ ll /mnt/afs/inria.fr/rocq/home/atoll/common/cvs Segmentation fault And I can't umount afs: Arrêt de AFS client daemon : umount: /mnt/afs: périphérique occupé umount: /mnt/afs: périphérique occupé The last point is a regression

[OpenAFS] unable to stop openafs

I'm unable to umount an afs partition: umount /mnt/afs fails because peripheral is still occupied, however I can't see any process using any file located under /mnt/afs using lsof. It seems however to be the canonical way to stop the afsd daemon, as afsd -shutdown exit when the afs partition is

Re: [OpenAFS] unable to stop openafs

Vladimir Nadvornik wrote: Hm. This is a problem we haven't been able to track down so far. Can you confirm that if you start AFS then read a file, you can unmount /afs, but if you do something like 'fs la' on the same file, you stop being able to unmount it? Yes, exactly. The same for me. -- When

Re: [OpenAFS] patch still needed for openafs 1.3.78

Derrick J Brashear wrote: On Mon, 7 Feb 2005, Guillaume Rousse wrote: The following patch is still needed for openafs 1.3.78, otherwise the current tree produced by make libafs_tree is not functional, as osi_vfs.h is generated in a wrong place. Does said patch still work when you're do

[OpenAFS] patch still needed for openafs 1.3.78

The following patch is still needed for openafs 1.3.78, otherwise the current tree produced by make libafs_tree is not functional, as osi_vfs.h is generated in a wrong place. -- A clean and dry set of overalls is a magnet for mud and rain -- Murphy's Bush Fire Brigade Laws n°16 -

[OpenAFS] unable to stop openafs

I'm unable to umount an afs partition: umount /mnt/afs fails because peripheral is still occupied, however I can't see any process using any file located under /mnt/afs using lsof. It seems however to be the canonical way to stop the afsd daemon, as afsd -shutdown exit when the afs partition is

Re: [OpenAFS] naive questions about openafs

Derrick J Brashear wrote: On Sat, 22 Jan 2005, Guillaume Rousse wrote: The point is: how would it change, otherwise as inside the source code ? I think I found an exemple with the rlim stuff in 2.6.10, that is handled by configure step :/ Right, this is what we've been trying to explain. Fo

Re: [OpenAFS] naive questions about openafs

Guillaume Rousse wrote: Matthew Miller wrote: On Fri, Jan 21, 2005 at 06:07:42PM +0100, Guillaume Rousse wrote: BTW, I could be wrong, but I don't really see kernel-tree-specific options in the generated makefile, apart build directory pathes that could get easily made relatives. I t

Re: [OpenAFS] naive questions about openafs

Derek Atkins wrote: Guillaume Rousse <[EMAIL PROTECTED]> writes: Right, but it still make symlinks to source files, generate the makefiles, and execute it immediatly with its own options. What I want is something ready to be included in current 2.6 linux tree, so as to be able to run i

Re: [OpenAFS] naive questions about openafs

Matthew Miller wrote: On Fri, Jan 21, 2005 at 06:07:42PM +0100, Guillaume Rousse wrote: BTW, I could be wrong, but I don't really see kernel-tree-specific options in the generated makefile, apart build directory pathes that could get easily made relatives. I tested against two differen

Re: [OpenAFS] naive questions about openafs

Derek Atkins wrote: Guillaume Rousse <[EMAIL PROTECTED]> writes: I don't care how much call to configure or make I have to do, what I want is to be able to reach a stage where I could invoke make with my own args to create a single kernel module. Current build system give me one ma

Re: [OpenAFS] naive questions about openafs

Derrick J Brashear wrote: On Fri, 21 Jan 2005, Guillaume Rousse wrote: I also tried the libafs_tree, and found a minor bug: the configure_libafs must be renamed to configure, otherwise the first configure test (ls -t) will fail. We can fix that. Anyway, it doesn't solve my problem, a

Re: [OpenAFS] naive questions about openafs

Matthew Miller wrote: I had a look at the rpm available from openafs website. However it seems to be uncompatible with kernel 2.6, so I'd have to package devel version instead, and it would be a lot of work anyway to adapt it to official mdk packaging policy. Look at my package at

[OpenAFS] naive questions about openafs

I'm interested about being able to use afs from mandrake, because it is widely used in my institution. However, I'm totally clueless about afs, so it's a bit difficult for me to figure what is needed exactly. First, current mdk kernel (2.6.10) ships an afs module. However, looking at source fil