On 07/20/2016 12:10 PM, Wols Lists wrote: > On 20/07/16 16:36, Wols Lists wrote: >> Get that first 3TB drive. NOW. Physically replace sda in the machine, >> and configure it as a single-drive mirror ( --create --devices=2 sda >> spare). > > Just noticed your edu address. If Phil Turmel chimes in, he'll tell me > off for telling you to spend money :-) No, not this time. :-) I would also note that the raid1 composed of the six sd?2 partitions is only operating as a two-copy mirror -- the other four devices are spares. Whoever created that array added the extra drives but never used --grow to set the number of mirrors to six. Amit, as soon as you get your arrays assembled with the --force option, follow Wol's advice on rsync'ing to a new array. I wouldn't advise trying to fix the UREs on your existing arrays since there are so many that you'd certainly trip the 10/hr read error limit in MD. Just copying data may also trip the read error rate limit, but you have no choice. If it happens and kicks out any array devices, simply stop the rsync, stop and re-assemble the array (with --force), and continue with rsync. The Constellations can probably be recycled into spares or backup devices after you copy your data. I didn't look close at their stats. Phil -- 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
