On Fri, 1 Apr 2005, Alvin Oga wrote: > > hi ya raiders .. > > we(they) have 14x 72GB scsi disks config'd as raid5, > ( no hot spare .. ) > > - if 1 disk dies, no problem ... ez to recover > > - my dumb question is, > - if 2 disks dies at the same time, i > assume the entire raid5 is basically hosed > if it won't reassemble and resync from > the point where it last was before the crash ?? It's possible to recover it - IF one of the failed disks hasn't really failed. ie. no genuine bad sectors or lost data. I had a 6-year old 8-disk array a while back that had been retired after 5 years of trouble-free operation, but was subsequently pressed into use on a different server - it featured some dodgyness about it - it would occasionally fail a disk because the sun was in the wrong place, or the moon was full, or something - never got to the bottom of it - the disks would always surface check OK afterwards - they may have been remapping sectors, but I never observed data or file system corruption, and I did occasionally get a 2-disk failure, but I was always able to resurect it using the last disk to fail as part of the array. Fortunately the stop-gap it was filling has been replaced by something new now! > - i assume that the similar 2 disk failure > also applies to hw raid controllrs, but it'd > be more dependent upon the raid controller's > firmware for it's ability to recover from > 2 of 14 simultaneous disk failures > ( lets say the dell powervault 2205 series ) > > - i think 4x 300GB ide disks is better ( less likely to fail ?? ) Who knows. With the H/W solution, you really are at the mercy of the hardware supporting software. Less disks might be less risk of failure though. Some modern disks don't seem to be having a good press recently though. (eg. Maxtor) I've switched to RAID-6 now, even for a 4-disk system I built recently. Disks are cheap enough now. (Unless you have to buy them from Dull or Stun!!!) > and yes it has already crashed twice with > 2 different disks running at 78F at nights > and weekends when the air conditioning is off Um - thats only 25C. Well inside the limits I'd have thought. I have some disks (Maxtors!) that are hapilly running at 50C (Although for how much longer, I don't know, but they have survived 15 months so-far, but they are in a fairly stable temperature environment - at the top of a lift shaft!) By comparison, I have another box (same config & age) thats effectively outside and the temperature cycles are very visible and it's just had a disk fail )-: Gordon - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
