Hi I summarized my readings from the mail list, hope that someone can correct me if I am wrong, thanks. 1. When md is degraded or during recoverying, read error is FATAL and non-recoverable. 2. After 2.6.16, RAID1 auto-correct read errors by re-write the block during resync or normal IO. 3. After 2.6.15, RAID5 auto-correct read errors by re-write the block on normal IO. 4. There is still one thing I do not know is that "How will RAID5 handle read errors during resync?" Thanks for your kind help : ) -- Best Regards, Michael Yao On 12/5/06, michael michael <danshoe@xxxxxxxxx> wrote:
Hi, I have several questions about RAID1/RAID5 read error handling. Any help is appreciated, thanks. After I created a md, what will md driver do if a read error is encountered in the following conditions: 1. during md is resyncing? 2. after md is resync-ed? 3. when md is degraded? 4. during md is recovering? My current kernel is 2.6.15.7, but I saw patches for the read errors handling in 2.6.16/18. What are the changes of read error handling after 2.6.16?
- 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
