[Linux-cluster] problems with iscsi storage and boot order

[Randy Brown]

I have a two node cluster configured which is going to be used as a NAS 
head for our iscsi based storage.  I think I have it working, for the 
most part (still some fencing issues, but that will be in another 
post).  I am using Centos 5 and it's associated clustering software.

If I boot both nodes with none of the clustering components (cman, 
clvmd, gfs, or rgmanager) starting at boot, I can restart the iscsi 
service then start cman, clvmd, gfs, and rgmanger and everything works.  
The problem stems from the iscsi not making a successful connection to 
the storage when the iscsi service starts so clvmd doesn't see the 
volume group so gfs can't mount the filesystem.

I have tried moving things around in the boot order using chkconfig, 
i.e., moved iscsi to S68, cman to S70, clvmd to S72, and gfs to S74 at 
runlevels 2, 3, 4, and 5, but that didn't make any difference.

I was wondering if anyone else has experienced this or has any 
suggestions as to what I can do.  It's very frustrating and I don't want 
to have to manually start the clustering on each machine after a reboot.

I will gladly provide further information as required.

Thanks in advance,

Randy