Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-07-14 Thread Steinar H. Gunderson
On Thu, Jun 14, 2007 at 05:18:38PM +0200, Bart Cortooms wrote: We're using drbd+heartbeat+nfs as well, and the symptoms are the same: when calling /etc/init.d/nfs-kernel-server stop from the command line all threads are stopped cleanly. When heartbeat calls /etc/init.d/nfs-kernel-server

Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-07-14 Thread Lars Ellenberg
On Sat, Jul 14, 2007 at 11:22:57AM +0200, Steinar H. Gunderson wrote: On Thu, Jun 14, 2007 at 05:18:38PM +0200, Bart Cortooms wrote: We're using drbd+heartbeat+nfs as well, and the symptoms are the same: when calling /etc/init.d/nfs-kernel-server stop from the command line all threads are

Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-07-14 Thread Steinar H. Gunderson
reassign 248300 linux-source-2.6.18 thanks On Sat, Jul 14, 2007 at 11:04:07PM +0200, Lars Ellenberg wrote: Could a reassign be appropriate? yes. IMO, it is likely some race condition in kernel nfs code. Then reassigning to 2.6.18. Thanks for the analysis. /* Steinar */ -- Homepage:

Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-06-20 Thread Michelle Konzack
Hello Bart and *, Am 2007-06-14 17:18:38, schrieb Bart Cortooms: We're using drbd+heartbeat+nfs as well, and the symptoms are the same: when calling /etc/init.d/nfs-kernel-server stop from the command line all threads are stopped cleanly. When heartbeat calls /

Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-06-14 Thread Lars Ellenberg
On Fri, Jun 08, 2007 at 12:20:12PM +0200, Steinar H. Gunderson wrote: On Thu, Jun 07, 2007 at 11:45:25PM +, Bart Cortooms wrote: /proc/nfs/nfsd does not exist (at least not on kernel 2.6.18). Instead the virtual filesystem for NFS is (now?) available at /proc/fs/nfsd. I replaced it

Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-06-14 Thread Steinar H. Gunderson
On Thu, Jun 14, 2007 at 04:46:25PM +0200, Lars Ellenberg wrote: we are not yet sure how to trigger the problem. if we try by hand it is not reproducable. if we try by heartbeat it happens all the time. That's an interesting data point. if that works for you, too, feel free to notify

Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-06-14 Thread Bart Cortooms
On Jun 14, 2007, at 5:11 PM, Steinar H. Gunderson wrote: On Thu, Jun 14, 2007 at 04:46:25PM +0200, Lars Ellenberg wrote: we are not yet sure how to trigger the problem. if we try by hand it is not reproducable. if we try by heartbeat it happens all the time. That's an interesting data point.

Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-06-08 Thread Steinar H. Gunderson
On Thu, Jun 07, 2007 at 11:45:25PM +, Bart Cortooms wrote: /proc/nfs/nfsd does not exist (at least not on kernel 2.6.18). Instead the virtual filesystem for NFS is (now?) available at /proc/fs/nfsd. I replaced it with the following: Thanks, I'll at least be putting in the patch. A pity it

Bug#248300: nfs-kernel-server: exportfs -f uses wrong path in init script

2007-06-07 Thread Bart Cortooms
Package: nfs-kernel-server Version: 1:1.1.0-3 Followup-For: Bug #248300 Hi, I stumbled on this old bug report while investigating a similar problem: sometimes umount fails with a device is busy after stopping nfs-kernel-server and nfs-common. After looking in /etc/init.d/nfs-kernel-server, it