On Tue, Aug 16, 2005 at 09:50:46AM -0700, Russ Allbery wrote: > Jose Calhariz <[EMAIL PROTECTED]> writes: > > > Package: openafs-modules-2.4.27-2-686-smp > > Version: 1.3.81-3+4 > > Severity: important > > > This month I started to test the stability of the openafs-client and > > openafs-modules on sarge with Debian kernel 2.4.27-2-686-smp. This week > > I have seen the clients applications blocked, openafs-client blocked, > > and two oops on different machines, and lost one volume because of > > corruption. This report is about one OOPS while doing a "tar xjf > > linux-2.6.9.tar.bz2" from a OpenAFS volume to the same OpenAFS volume, > > while other machines stressing the same volume. The tar seams blocked > > until now. > > 1.3.87 has now been uploaded to Debian unstable; if you could give a > kernel module built from that openafs-modules-source package a shot, I'd > be very interested to see if the results are any different. >
I have now one machine that is compiling the linux kernel, on an openafs volume, for 24 hours. Until now I didn't get any error from the compilation or message from the kernel. That is very good and much better than with 1.3.81/2 modules. This client have openafs-client from sarge (1.3.81) and openafs-modules from unstable (1.3.87) > This does sound like a good test case; I'll give it a shot when I have a > chance (I have a bunch of other packaging work that I need to do first, > unfortunately). > José Calhariz -- Toda a história do mundo se resume ao fato de que quando as Nações são fortes nem sempre são justas, e quando elas querem ser justas não são suficientemente fortes --Winston Churchill
signature.asc
Description: Digital signature