Johan, > > WD has had problems similar to this with many of their drives. It just > > decides to 'go away'. There is a fix available on their web site for the > > 180GB and 200GB drives (and a better description of the problem), but the > > problem is NOT limited to those drives. > > How do these problem appear in log files? > > I have a machine with two Promise Ultra100 TX2 cards, and five > WD2000JB 200 GB drives in RAID-5. In a month, i've had a few disk "failures" > that typically looks like this in the logs: <SNIP LOG> > The disk itself doesn't appear to know about any failures > (using smartctl), and it works again when hotadded to the raidset. I've > also had a multiple drive "failure" twice, both times with two drives > using the same IDE channel. > > I'm not sure if these problems are caused by buggy Promise ATA drivers > in my kernel (RH9, 2.4.20) or the WDC problem with 180/200 GB drives. > From WDC's description of the problem, I got the impression that it > only happened when the drives were connected to hardware RAID cards > like 3Ware IDE raid controllers. This appears to be the WD problem. It is caused by some timing-related irregularities in their microcode. It occurs in a 'RAID environment'. The article is not specific about only hardware RAID cards. There is a fix package for 3Ware cards and a fix package for non-3Ware cards. Use the package for non-3Ware cards. If you need to have more than 4 drives in an array, with 2 Ultra100 cards, you might want to consider adding a Promise Ultra133 card. You are normally limited to 2 of each, for a total of 4 cards (8 channels total). This would solve your double-failure problem. FYI, I'm running the reverse of this (2 Ultra133 + 1 Ultra100) in one of my file servers with excellent results. Good luck. Peter Ashford - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
