Quote:
Originally Posted by nepenthes9
redhat cluster suite on CentOS 5.2, following is my cluster.conf:
<< BIG SNIP >>
<resources>
<clusterfs device="/dev/sdb5" force_fsck="0" force_unmount="1" fsid="45793" fstype="ext3" mountpoint="/nfsdata" name="nfsdata" options="rw"/>
<nfsclient name="any_host" options="rw,async,no_root_squash" path="" target="192.168.1.0/24"/>
<nfsexport name="nfs_export"/>
<ip address="192.168.1.50" monitor_link="1"/>
</resources>
<service autostart="1" domain="nfsha" name="nfs1" nfslock="1">
<ip ref="192.168.1.50"/><clusterfs ref="nfsdata">
<nfsexport ref="nfs_export"><nfsclient ref="any_host"/></nfsexport>
and my log:
Aug 29 12:31:39 tpe52 clurgmgrd[19155]: <notice> Starting disabled service service:nfs1
Aug 29 12:31:39 tpe52 clurgmgrd: [19155]: <err> No export path specified.
Aug 29 12:31:39 tpe52 clurgmgrd[19155]: <notice> start on nfsclient "any_host" returned 2 (invalid argument(s))
Aug 29 12:31:39 tpe52 clurgmgrd[19155]: <warning> #68: Failed to start service:nfs1; return value: 1
Aug 29 12:31:39 tpe52 clurgmgrd[19155]: <notice> Stopping service service:nfs1
Aug 29 12:31:39 tpe52 clurgmgrd: [19155]: <err> No export path specified.
Aug 29 12:31:39 tpe52 clurgmgrd[19155]: <notice> stop on nfsclient "any_host" returned 2 (invalid argument(s))
Aug 29 12:31:40 tpe52 clurgmgrd[19155]: <crit> #12: RG service:nfs1 failed to stop; intervention required
Aug 29 12:31:40 tpe52 clurgmgrd[19155]: <notice> Service service:nfs1 is failed
Aug 29 12:31:40 tpe52 clurgmgrd[19155]: <crit> #13: Service service:nfs1 failed to stop cleanly
failed.
|
--------
<nfsclient name="any_host" options="rw,async,no_root_squash" path="" target="192.168.1.0/24"/>
I strongly suspect your problem(s) [as indictated by the error log entry above] is with the way you have configured the NFS service.
I have space out the error log so that you can see the reported issues clearly. XML is simple, each section is simply surronded by labels that start like: <resources> and always finish by the "/" preamble, as in </resources>. Seperating and spreading the xml configuration out helps you better see that is involved.
There are several issues reported, but you should always start with the first reported problem because most times, clearing the first issue resolves other outstanding errors.
So, given that the first reported error is: nfsclient adress is unresolvable. This is because the target address given is a network address rather than a specific machine address.
===/===
IMHO is best you learn about NFS first, so then you can better understand your clustering configuration.
Read "man nfs -k" information and simply set up a NFS share [looking and using all the tools used to ensure correct setup eg: rpcinfo -p, mount, showmount etc.] between a two machines.
Once your got this simple NFS setup configured, up and running, then you will know how to set the parameters correctly for what you need. Now go back to the cluser configuration and correct the NFS setup to the target your trying to use.
You can used the newly learnt tools to debug where-else you may have issues.
Hope this assists you.