Hi there, I'm currently implementing the NFS/DRBD how-to on debian etch, but have come up against a problem with step 7, the moving of the nfs directory from /var/lib/nfs to /data/nfs. when I try and delete or move the directory I get the following errors: rm: cannot remove directory `/var/lib/nfs/rpc_pipefs/statd': Operation not permitted rm: cannot remove directory `/var/lib/nfs/rpc_pipefs/portmap': Operation not permitted rm: cannot remove directory `/var/lib/nfs/rpc_pipefs/nfs': Operation not permitted rm: cannot remove directory `/var/lib/nfs/rpc_pipefs/mount': Operation not permitted rm: cannot remove directory `/var/lib/nfs/rpc_pipefs/lockd': Operation not permitted I have tried shutting down anything that might be locking those files by doing things like: /etc/init.d/nfs-kernel-server stop /etc/init.d/nfs-common stop killall -9 portmap but it still won't let me do it. Any help or suggestions would be greatly appreciated! Thanks Simon
Ok, I scrapped everything and started again... and this time it just worked, so no idea what the problem was first time around! Thank you Falko for the great tutorials! I'm now going through the HA Apache Cluster How-To. I've completed this successfully before as originally we were looking at a 6 machine cluster, but to save power costs we are now looking at reducing that to 4 by putting the Loadbalancers on the same machines as DRBD. Do you forsee any issues with doing this? And if not, could you tell me how I go about combining the heartbeat configurations from the two articles so that nothing conflicts? Ideally I would like one machine to be the master LB and the other to be the master DRBD to spread resources the most efficiently, with each returning to it's original server after a failure is resolved. Thanks again, Simon
Can anyone give me any pointers on combining the two ha.cf's? I think for the way I want to do it all I need to do with the haresources file is something like: ... assuming of course that I can have more than one virtual IP on a single pair of servers. What I need to understand is why each ha.cf is how it is... so for the HA-NFS we have: And for the HA-Apache we have: Now, assuming the one and two line "node" statements are just a different way of stating the same thing, the common lines are: For "NFS" we have the following additional lines: And for "Apache" we have these: So my questions are: 1. Is there a specific reason why you would NOT want to use auto_failback for the Load Balancers, as configured in the How-To? 2. If I put everything into one ha.cf will the lines that were not specified in the ha.cf for one service have an adverse effect on the running/stability of the other? 3. Can anyone explain to me what those 6 uncommon ha.cf lines are there for in relation to the how-to that they came from? Falko's how-to's have been absolutely fantastic for going through step by step and having it just work, but now I need a bit more understanding of WHY it's configured how it is.
Vi /etc/exports Path_share IP_Share(permission) If you want to share All IP = * open port 111 + range port 500-1000 in firewall Use user roor
