[OpenAFS] SuSe is frustrating
Fri, 27 May 2005 12:11:46 -0500
>On Fri, May 27, 2005 at 10:36:03AM -0400, Randy Lee Schott wrote:
>>> I'm not even sure I'm posting to the right list, but I figure it's worth a
>>> So I'm trying to use openAFS 1.3.78 under SuSe 9.3 (Linux 2.6 kernel).
>>> Believe me, I would've used a stable release if I could have. So, I got it
>>> to build for my kernel, I can start the afs service (afs start) with no
>>> problems/error messages. I can start afsd as well with no problems or
>>> error messages. I can even klog to my cell to get tokens. When I try and
>I am using the packaged openafs that suse put in 9.3. It is openafs
>1.3.78... I had to make some changes the suses configuration scripts,
>but it all works fine. I am using the smp kernel as well.
>22.214.171.124-20a-smp on a dell precision dual xeon box.
>I attempted to compile openafs 1.3.8x, but if the make fails beyond
>the ./configure, make, make install, I don't know enough to fix
>things. Well, the build failed. Which is why I used the suse openafs
>rpms. It was a big enough pain to fix their scripts.
>why not just use the suse rpms since they are the same version you
I've been getting decent results on SuSE 9.3 running the CVS version of the
OpenAFS kernel module with the SuSE 9.3 supplied rpms.
I set the kernel in /usr/src/linux up to build against by doing a
cp /boot/symvers-126.96.36.199-20a-i386-default.gz /usr/src/linux/Module.symver.gz
I build the kernel module from CVS by doing a CVS update, then running
./configure --with-linux-kernel-headers=/usr/src/linux --sysconfdir=/etc/openafs
Then I copy the libafs ko file into /lib/modules/'uname -r`/extra/libafs.ko
I'm curious about what you had to change in the SuSE init scripts? The
only changes I had to
make to configure were in /etc/sysconfig/afs-client, where I set it to
not regenerate cell info
and to use the disk cache. So, I didn't have to change the scripts at all.
>>> cd to my personal afs space (cd /afs/my.cell/myspace) or even when I try
>>> and browse my root directory in Konqueror, the system hangs. I'm unable to
>>> kill any afs related processes(kill -9 pid) and have to restart the machine
>>> to fix it. I have my firewall set to allow udp ports 7000-7009(afs
>>> services) 1024-2048(I'm told klog needs these), and 123/tcp(told that ntp
>>> uses it). My ThisCell file contains only my cell, and my CellServDB file
>>> has been reduced to just the info for that cell, so it shouldn't even be
>>> trying other cells, correct? I've been browsing archived mailing lists and
>>> forums for days now trying everything I can, but I'm just not sure what is
>>> causing this problem. Has anyone else seen this? Thanks a lot.
>>> OpenAFS-info mailing list
> -- David Bear phone: 480-965-8257 fax: 480-965-9189 College of Public
> Programs/ASU Wilson Hall 232 Tempe, AZ 85287-0803 "Beware the IP
> portfolio, everyone will be suspect of trespassing"
> OpenAFS-info mailing list OpenAFSfirstname.lastname@example.org