Because of Error in the Red Hat Cluster Suite I have to work at Saturday. The Cluster Suite sucks really in the RHEL4 U4 Version a lot of critical Errors pops up: rgmanager memoryleak, rgmanager segfault, dlm problems and also dlm timeouts because there is no effective GFS lock purge mechanism in U4. We have min. 32 GB RAM and U4 can’t handle so many GFS locks that can be created with this memory size, at least in our Config 😉 . We allways hiting the bug first. Also we use open-sharedroot.org extension and in this Case a GFS – freeze is really bad because then debugging is very painful.
Also every time when i look very deep at the system I find something. This time I find a Hardware error on another Cluster and we forgot in the test environment a important packages ( we use the open-sharedroot.org extension and in this case you must link /proc/mount to /etc/mtab. And for this the mount command need a patch to NOT mount the bind Filesystems twice! )
Also the cp of the NFS Filesystem goes 3 x times as expected. All in all I missed my Lunch with the Family, but next Friday I take a day off and go earlyer to the Salwideli.org