May 12, 200917 yr But whats the ultimate goal... maintain the ability to not lose data as close to 100% of the time as possible. I agree. But we have ways to do almost anything without losing parity protection ... 1 - Adding a drive (using preclear script) 2 - Removing a drive (see this thread) 3 - Replacing a drive (you just need one extra port to be able to add the new drive, copy data from old drive to new drive, and finally remove the old drive - as bubbaQ described) The only things that can't be done is rebuild a disk, and this is only needed when an actual failure occurs. Should be very rare. The only time you are not protected from a disk failure is when you have already had a disk failure and are in the process of rebuilding the failed disk.
May 13, 200917 yr The only things that can't be done is rebuild a disk, and this is only needed when an actual failure occurs. Should be very rare. I thought you had the ability to rebuild one disk by re-assigning it on the devices page.
May 13, 200917 yr The only things that can't be done is rebuild a disk, and this is only needed when an actual failure occurs. Should be very rare. I thought you had the ability to rebuild one disk by re-assigning it on the devices page. You cannot rebuild a disk while still maintaining parity protection. If you were to have another disk fail while rebuilding a disk, you would lose data. All of the other situations - if you follow the safe procedures - would allow you to recover from a single failure. I have to take back (a little) the comment about losing data while rebuilding a disk. If you were to "upsize" a disk, and lose a disk while upsizing - you could recover and not lose data if you kept the original disk unaltered during the rebuild process.
May 13, 200917 yr Hope this makes more sense than my mumblings from last night While you are rebuilding a failed disk you are not protected from a (second) disk failure. But in normal growing/upsizing/removal situations, there are safe procedures to stay parity protected throughout. In some cases the safe procedures take longer and have more steps to follow, but for some this is a small price to pay for continuous parity protection. For example, if you want to remove a disk from the array, you have 2 options. THe "unsafe" way is to unassign the disk on the devices tab, press restore, start the array, and unRAID will rebuild parity (a long process during which you are NOT parity protected). The "safe" alternative is to fill the disk you want to remove with binary zeros while it is still part of the array (an even longer process BUT during which you ARE parity protected). After that, with a few quick steps, you can unassign the disk, start the array, and be instantly parity protected. (I posted a link to a detailed procedure a few posts back in this thread). Am I forgetting something? Are there routine tasks that cause you to lose parity protection for long periods of time?
Archived
This topic is now archived and is closed to further replies.