Hi all, I just removed the use of the (borrowed from memexp) cidev when using the OpenDLM lock module. Now, the configuration information (i.e. journal ID, # nodes in configuration) comes from calls into OpenDLM itself, both calls in src/kernel/dlmdk/dlm_base.c: clm_getlocalid() clm_node_count() This info originates from the OpenDLM configuration file, /etc/dlm.conf. As before, there is an offset of 1 between the OpenDLM nodeid (range of 1 - n), and the OpenGFS "jid" journal ID (range of 0 - (n-1)). I have successfully mounted OpenGFS/OpenDLM on one node, and have done a read and write (using vi). A second node still won't mount, so we're still working on that problem. One thing that I noticed is that it takes a bit of time for OpenDLM to get its own cluster configuration together when firing it up on just one node (OpenDLM node 2), and so the mount of OGFS will fail if I don't wait long enough. It aborts cleanly, however, and I can try again successfully a few seconds later. I've checked in the code ... I'll soon update the HOWTO to eliminate the step of creating the cidev for OpenDLM. BTW, this required using an addition OpenDLM include file, dlm_kernel.h. This file had redundant (when building with OpenGFS) definitions of "TRUE" and "FALSE". I wrapped these with #ifndef so it would build cleanly (and checked it into OpenDLM CVS). SO, get the latest CVS from both OpenDLM and OpenGFS if you want to try this out. -- Ben -- ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id56&alloc_id438&opÌk _______________________________________________ Opengfs-devel mailing list Opengfs-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/opengfs-devel