On Thu, Dec 2, 2010 at 2:17 PM, Spelic <spelic@xxxxxxxxxxxxx> wrote: > It's already applied only in 2.6.37-rc1..rc4 After one year of mvsas troubles, I can finally say that 2.6.37-rc4 fixed all of my mvsas issues. I have been stress testing 2 machines for days and can't trigger the usual mvsas issues (nexus errors and unplug notices) anymore. In the past, on one system all 8 1TB drives of 2 different brands (Hitachi & Samsung) on a Supermicro AOC-SASLP-MV8 gave me random errors, such as unplug notices. Large file copies over samba stalled or were aborted. This could easily be triggered by doing 1 hour of large IO. On a second system, newer 1.5 TB drives gave me almost no issues. I had to stress much longer before something appeared in the log. Both machines have identical motherboards and same kernel versions. So my suspicion some of the older 1 TB drives were bad is now confirmed. With 2.6.37-rc4, only the real faulty drives are now reported in dmesg instead of ALL drives after waiting long enough. I found one drive had 12 uncorrectable sectors according to smart, while another drive always reported DriveStatusError under heavy load. These two drives were always slowing down the array. After removong them from the RAID6 software raid, speed is now stable. On the second system with all good drives, the nexus errors and unplug notices are completely gone. Thanks to everyone who helped making mvsas stable as it finally works! Best regards, Audio Haven (Frederic Vanden Poel) -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
