I think you are looking at something more like this : https://www.google.co.uk/imgres?imgurl=https%3A%2F%2Fthumbs.dreamstime.com%2Fz%2Fhard-drive-being-destroyed-hammer-16668693.jpg&imgrefurl=https%3A%2F%2Fwww.dreamstime.com%2Fstock-photos-hard-drive-being-destroyed-hammer-image16668693&docid=Ofi7hHnUFmPsyM&tbnid=Ak6YfqQVvZWCsM%3A&vet=10ahUKEwj56JfI5vrVAhXoCcAKHfkZDn4QMwgmKAAwAA..i&w=1300&h=1130&client=safari&bih=1116&biw=1920&q=hdd%20hammer&ved=0ahUKEwj56JfI5vrVAhXoCcAKHfkZDn4QMwgmKAAwAA&iact=mrc&uact=8
:P I’ll sound brusk now so buckle up. You really need to set your priorities straight now, if you want to rescue a disk that has a pending sector you set your self up for failure. You said that several consecutive power outage killed your cluster, yet you did show no concern of investing in at least anti surge protector (outages can create surges, or a surge can cut the power by burning fuses in sub station) which actually cause hardware failures. I’ve fired at you a control statement about how to save your data, but you keep returning trying to save some osd’s - which to me look like you don’t really care about your data. If any of those were at any point on your mind, you would shut down those systems to limit possibility of another outage destroy more data. You would protect your self via a simple surge protector or at least basic UPS. You would borrow / beg / steal to get a spare hard drive and backup your stuff. You would set your pool size and min size to 1, let the cluster get out of warning state and only then you will be able to mount it to attempt data recovery. You would check remaining disks for SMART errors. Then and only then you can start playing with as complex repairs. Right now you are in the middle of a shit creek without a paddle. _______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
