You go after these things, until you realize you are making no progress, and are not going to, without a break. As of this moment, there is no conclusive fix, but a couple of things are now known.
It's not Rivendell related. That's definitive. It is a CentOS / RHEL / GNOME thing. It has reared it's ugly head from time to time in various versions of RHEL going back some ten years, possibly more. It has also reared it's ugly head in various incantations of Ubuntu. It appears to be NFS related, but we're not certain. If the client is CentOS 6.8, and the server is CentOS 6.5 the problem does not manifest. An identical server but CentOS 6.4 the problem does manifest. If the client is CentOS 6.7 the problem does not appear at all. Once the NFS share is mounted, gam_server holds it. It cannot be umounted without hacking gam_server. If NFS is never mounted, the problem doesn't manifest. ( at least in our test cases ) I have a few more thoughts, but can't test them until Monday or so. It also appears that while gam_server will consume 100% of cpu when nothing else is requesting cpu, it's really more of an annoyance to those of us who bother to look. It appears to not affect the performance of Rivendell at all. Not by itself, anyway. -- Cowboy http://cowboy.cwf1.com Insanity is the final defense ... It's hard to get a refund when the salesman is sniffing your crotch and baying at the moon. _______________________________________________ Rivendell-dev mailing list [email protected] http://caspian.paravelsystems.com/mailman/listinfo/rivendell-dev
