Hello, I discovered a workaround using an undocumented feature of GFS and gnbd. (did not found it at RH manuals and man pages) Maybe it does not happen in red hat cluster source newer than 1.03.00, kernel 2.6.18, when used with high performance hardware and hi perf private network, but the problem is clear at low end hardware and networking. the code could be more tolerant to these. Described the work around at [0] Regards. Andre Felipe Machado [0] http://www.techforce.com.br/index.php/news/linux_blog/red_hat_cluster_suite_debian_etch -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster