I have concluded several rounds of stress testing on 2.4.0 with Maxtor and IBM drives, fast with optimisation set to the highest, up to 30 routines writing to disk running in parallel. I am unable to reproduce the error you reported. Perhaps some additional information on your hardware is appropriate. lspcidrake > file dmesg >> file cat /etc/lilo.conf >> file cat /etc/modules.conf >> file then send me file attached to an email Civileme -- QA/Software Testing