KR> We don't use this, which is fine. However, leads to a KR> question.. why the decision to have do_snapshot() return void? Because the status and state information is conveyed through the job. What else could the caller do with a return value? It's in a thread. KR> The function is defined as vss_get_save_path() here, but called as KR> get_save_path() later on. Wow, that was dumb. I changed this right before I sent it out, but I think it compiled the first time because I had the function defined in the installed libraries and headers. Thanks :) KR> I assume this path will be valid on all systems / distros? Probably not. We probably need to put this in our configure.ac. KR> Need to initialize s here. Or make sure to set the status when KR> the calloc fails in new_context(). Yep, thanks. -- Dan Smith IBM Linux Technology Center Open Hypervisor Team email: danms@us.ibm.com