Hi Carlos: > What we can observe is that communications through Intel Gigabit cards > in the PCI slot is extremely slow, due to timeouts. Messages of this > kind, > e1000: eth0: e1000_clean_tx_irq: Detected Tx Unit Hang > are obtained from the kernel.
This is on the headnode? Or on the node's console during imaging? > We have done some trial with other OS, like FC4_64, SL4_64, FC3_32, > but without any success. 3Com Fast-Ethernet cards work fine, but are > slow for our purposes. > Now I am waiting for new 3Com Gigabit cards, or any other solution > that our IT dept. provides. > > I hope to resume the Oscar 4.2.1 work, and get the new > cluster operating soon. Are you by any chance using pfilter? You might consider upgrading the kernel on your headnode (via yum is fine) and see if it helps. BTW, will you be interested in helping us support Fedora Core 4 x86_64 for OSCAR 5.0? Thanks, Bernard ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid3432&bid#0486&dat1642 _______________________________________________ Oscar-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/oscar-devel
