![]() Once few bad blocks develop they may stay that way or may develop into a catastrophic disk failure in a week. If you see something other than 0 - RMA the drive.īy the way, synology default threshold for bad sector count before it complains is 50. However that kind of severe instability would have manifested itself during sync as well, and would have likely resulted in disk firmware relocating those faulty sectors.Ĭheck SMART data for relocated sector count on each drive. If you suspect that the bit rot has “crept in” during the couple of days since the sync - well, you have got defective drives. ![]() ![]() This is below the filesystem in this case but it does not matter: right after sync the data written on disks and raid arrays understanding of what’s written are in complete harmony. After raid sync all data is, well, “in sync”. Raid sync results in reads and writes all over the place, to compute parity and write it distributed across the drives. You have to balance it with how often can you tolerate performance degradation of your array during scrub. So what’s a good ballpark? Once every three months is OK if not too often. And then go to Vegas, apparently you are on a lucky streak :) But even if this does happen (you won the lottery) - remember - you did not touch that data for a while so that data has been backed up already and therefore you can restore corrupted block. Avoid a very unlucky situation when data rots on the same sector on more drives than fault tolerance of your array.To detect block on HDD that went bad and RMA HDD while it is still under warranty.So why run scrub at all then? To periodically check sectors that you haven’t touched for a long while for rotting for the following main reasons: This is the same process that happens when you read data normally - if the data you tried to read happens to be corrupted btrfs will recover it using available redundancy and return corrected data. If any of the data “copies” (lets leave it at this for simplicity) is not matching as a result of a bit-rot the filesystem will repair the invalid sector based on checksums and return correct data. Long answer: Scrub reads all data from the raid and verifies data consistency. Short answer - it does not matter as long as you do schedule it to run at least maybe annually.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |