[gentoo-user] nfs server problem
Hello, I installed a vanilla 2.6.20 kernel in order to (eventually) run kerrighed. The kernel boots fine but the nfs server won't start and I see this in the logs: Oct 3 07:34:40 lowalbite rpc.statd[103835]: Version 1.1.0 Starting Oct 3 07:34:41 lowalbite nfsd: last server has exited Oct 3 07:34:41 lowalbite nfsd: unexporting all filesystems Oct 3 07:34:41 lowalbite nfsd[103901]: nfssvc: Address already in use When I boot into gentoo 2.6.22 this problem goes away, so I assume there is some difference in kernel configuration. I have not been able to find what that is by comparing the NFS_* entries in the respective .congig files: they are identical. Can someone suggest how I might track this down? Thanks, Roger -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] nfs server problem
On Wed, 03 Oct 2007 08:54:30 -0230 Roger Mason [EMAIL PROTECTED] wrote: Hello, I installed a vanilla 2.6.20 kernel in order to (eventually) run kerrighed. The kernel boots fine but the nfs server won't start and I see this in the logs: Oct 3 07:34:40 lowalbite rpc.statd[103835]: Version 1.1.0 Starting Oct 3 07:34:41 lowalbite nfsd: last server has exited Oct 3 07:34:41 lowalbite nfsd: unexporting all filesystems Oct 3 07:34:41 lowalbite nfsd[103901]: nfssvc: Address already in use When I boot into gentoo 2.6.22 this problem goes away, so I assume there is some difference in kernel configuration. I have not been able to find what that is by comparing the NFS_* entries in the respective .congig files: they are identical. Can someone suggest how I might track this down? Thanks, Roger It seems like there's two options: either you're trying to run 2 nfs servers, or you're trying to run nfs threads that for some reason conflict with each other. anyway, the solution I suggest is checking the output of 'netstat -l -p -n' to see whether anything really is listening on those ports. If not, try using a quick script to keep reading the output of the previous netstat command and checking consistantly to see whether anything's listening on those ports while you restart nfs. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] nfs server problem
Hi Dan, Dan Farrell [EMAIL PROTECTED] writes: On Wed, 03 Oct 2007 08:54:30 -0230 Roger Mason [EMAIL PROTECTED] wrote: Hello, I installed a vanilla 2.6.20 kernel in order to (eventually) run kerrighed. The kernel boots fine but the nfs server won't start and I see this in the logs: Oct 3 07:34:40 lowalbite rpc.statd[103835]: Version 1.1.0 Starting Oct 3 07:34:41 lowalbite nfsd: last server has exited Oct 3 07:34:41 lowalbite nfsd: unexporting all filesystems Oct 3 07:34:41 lowalbite nfsd[103901]: nfssvc: Address already in use It seems like there's two options: either you're trying to run 2 nfs servers, or you're trying to run nfs threads that for some reason conflict with each other. anyway, the solution I suggest is checking the output of 'netstat -l -p -n' to see whether anything really is listening on those ports. If not, try using a quick script to keep reading the output of the previous netstat command and checking consistantly to see whether anything's listening on those ports while you restart nfs. -- That is agood idea -- I'll try it tomorrow. In the meantime I am working round it by setting the port(s) in /etc/conf.d/nfs, but it would certainly be cleaner to find and eliminated the conflict. Thanks for your help. Roger -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] nfs server problem
The error you are receiving indicates that you already have a service occupying the port that the nfs server wants to bind to, however only one service can bind to a port at any one given time. So you must find the service that is occupying the port nfs server wants to bind to and either disable that service or configure it to run on another port.
