You are not logged in.
I'm using nfsv4 to share between nodes on my LAN. Usually everything works great, but in the scenario of the nfs server getting shutoff, the nodes come unresponsive if a program attempts to access the now defunct NFS share. Is there anything I can do to change this behavior?
For example, if I attempt to umount the share on the node, even a "force lazy" the process just hangs; if I try to reboot the node, it hangs on trying to umount file systems. I found that I can comment out the corresponding entries in /etc/mtab and then reboot, but this is a manual process. I'd like it to be more robust.
Last edited by graysky (2011-06-22 07:43:21)
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
I understand what you mean but I'm not sure what causes it.
Have you tried mounting as both hard and soft to see if either makes a difference? Also I'm wondering if there is a timeout period where it just gives up.
Offline
Are you using the 'soft' mount option?
EDIT: Too slow
Last edited by fukawi2 (2011-06-22 08:23:29)
Are you familiar with our Forum Rules, and How To Ask Questions The Smart Way?
BlueHackers // fscanary // resticctl
Offline
I'm using nfsv4 to share between nodes on my LAN. Usually everything works great, but in the scenario of the nfs server getting shutoff, the nodes come unresponsive if a program attempts to access the now defunct NFS share. Is there anything I can do to change this behavior?
For example, if I attempt to umount the share on the node, even a "force lazy" the process just hangs; if I try to reboot the node, it hangs on trying to umount file systems. I found that I can comment out the corresponding entries in /etc/mtab and then reboot, but this is a manual process. I'd like it to be more robust.
Use automount/autofs
My blog: blog.marcdeop.com
Jabber ID: damnshock@jabber.org
Offline
Hi guys - just found out about the soft mount option for nfs thanks to your post. Since I specified no option, they were hard mounts. I have since modified them to be soft but the problem persists.
1) I stop the nfs server on the server
2) I issue a shutdown to the node but the process hangs
$ mount|grep work
mars:/work on /media/work type nfs (rw,soft,vers=4,addr=192.168.1.12,clientaddr=192.168.1.13)
Last edited by graysky (2011-06-22 21:13:19)
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
I have since modified them to be soft but the problem persists.
FWIW, I see the same issue during shutdown with my mounts if they fail, but they _do_ timeout after a couple of minutes... I only reboot once every month or 2 so I haven't bothered to try making it timeout faster.
Are you familiar with our Forum Rules, and How To Ask Questions The Smart Way?
BlueHackers // fscanary // resticctl
Offline