![]() |
Home
| Calendar
| Mail Lists
| List Archives
| Desktop SIG
| Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings Linux Cafe | Meeting Notes | Linux Links | Bling | About BLU |
On 01/27/2009 04:15 PM, Nuno Sucena Almeida wrote: > Hi Jerry,=20 > > I remember having locking problems with one of the rpc daemon about a > year ago, maybe stat, that would cause havoc with the nfs mounted > clients. When I looked at it, I traced to a flaw in some of the kernel > code which was corrected in some newer kernel version. Right now I'm > running 2.6.24 or above. I don't know how difficult is to try a new > kernel on your system, but maybe you can give it a try. > =20 I have a bit more information. I was able to find the lines of code, and = it has nothing to do with mmap(2). The problem is with fcntl, and only=20 when the NFS server is RHEL 5.2 Kernel 2.6.18-92. Works fine on a local=20 file (eg. /tmp or on a file served by RHEL 4.6 kernel 2.6.9-67. I handed = the problem off to our IT person to file a bug on RHEL 5.2. I should be=20 able to get the latest kernel for my NFS server, but it might be=20 difficult to get the time to reboot while people are doing work and have = some processes that run for hours or days (CPU bound, not I/O bound). fd =3D open(fileName, O_RDWR | O_CREAT, 0644); fl.l_type =3D F_WRLCK; fl.l_start =3D 0; fl.l_whence =3D SEEK_SET; fl.l_len =3D 0; fl.l_pid =3D getpid(); int fcnt =3D fcntl(fd, F_SETLK, &fl); /* fails when NFS server is RHEL = 5.2 */ --=20 Jerry Feldman <gaf-mNDKBlG2WHs at public.gmane.org> Boston Linux and Unix PGP key id: 537C5846 PGP Key fingerprint: 3D1B 8377 A3C0 A5F2 ECBB CA3B 4607 4319 537C 5846
![]() |
|
BLU is a member of BostonUserGroups | |
We also thank MIT for the use of their facilities. |