华为计算微信公众号
华为计算微博
华为计算今日头条
我的osd服务节点上有10块7.3T的机械硬盘和3块7T的nvme盘。在我创建集群的时候，我把其中两块nvme盘分别做成了10个逻辑卷供osd做为wal和db使用。
nvme0n1 259:0 0 7T 0 disk ├─vg--osd--wal-lv--osd--walk 253:20 0 715.4G 0 lvm ├─vg--osd--wal-lv--osd--walj 253:19 0 715.4G 0 lvm ├─vg--osd--wal-lv--osd--walh 253:17 0 715.4G 0 lvm ├─vg--osd--wal-lv--osd--walf 253:15 0 715.4G 0 lvm ├─vg--osd--wal-lv--osd--wald 253:13 0 715.4G 0 lvm ├─vg--osd--wal-lv--osd--wall 253:21 0 715.4G 0 lvm ├─vg--osd--wal-lv--osd--wali 253:18 0 715.4G 0 lvm ├─vg--osd--wal-lv--osd--walg 253:16 0 715.4G 0 lvm ├─vg--osd--wal-lv--osd--wale 253:14 0 715.4G 0 lvm └─vg--osd--wal-lv--osd--walc 253:12 0 715.4G 0 lvm
nvme1n1 259:1 0 7T 0 disk ├─vg--osd--db-lv--osd--dbk 253:10 0 715.4G 0 lvm ├─vg--osd--db-lv--osd--dbi 253:8 0 715.4G 0 lvm ├─vg--osd--db-lv--osd--dbg 253:6 0 715.4G 0 lvm ├─vg--osd--db-lv--osd--dbe 253:4 0 715.4G 0 lvm ├─vg--osd--db-lv--osd--dbc 253:2 0 715.4G 0 lvm ├─vg--osd--db-lv--osd--dbl 253:11 0 715.4G 0 lvm ├─vg--osd--db-lv--osd--dbj 253:9 0 715.4G 0 lvm ├─vg--osd--db-lv--osd--dbh 253:7 0 715.4G 0 lvm ├─vg--osd--db-lv--osd--dbf 253:5 0 715.4G 0 lvm └─vg--osd--db-lv--osd--dbd 253:3 0 715.4G 0 lvm
我现在已经成功的创建了第一个osd，但是我开始怀疑它的合理性。假如我的一块nvme坏了的话，那么我这个节点上所有的osd都会失去wal或者db数据。ceph集群是否可以应对这种情况？或者我应该如何规划磁盘的使用才更加的合理？
====== osd.0 ======= [block] /dev/ceph-22fe0799-26b4-4e22-b28b-8eb1d4d0c42d/osd-block-1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 block device /dev/ceph-22fe0799-26b4-4e22-b28b-8eb1d4d0c42d/osd-block-1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 block uuid wMlS8p-nMeS-xRXm-R2fe-1kFM-MIYM-JWi2ie cephx lockbox secret cluster fsid ce5fe6a9-f807-4198-b8d4-c284322c82fa cluster name ceph crush device class None db device /dev/vg-osd-db/lv-osd-dbc db uuid OtPXxw-Oh9L-Mcur-tnTU-sEzN-k6Kk-uoir86 encrypted 0 osd fsid 1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 osd id 0 osdspec affinity type block vdo 0 wal device /dev/vg-osd-wal/lv-osd-walc wal uuid JeSjju-3Lvc-3rCW-7dCH-44f3-2zAb-CPJyQG devices /dev/sdc [db] /dev/vg-osd-db/lv-osd-dbc block device /dev/ceph-22fe0799-26b4-4e22-b28b-8eb1d4d0c42d/osd-block-1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 block uuid wMlS8p-nMeS-xRXm-R2fe-1kFM-MIYM-JWi2ie cephx lockbox secret cluster fsid ce5fe6a9-f807-4198-b8d4-c284322c82fa cluster name ceph crush device class None db device /dev/vg-osd-db/lv-osd-dbc db uuid OtPXxw-Oh9L-Mcur-tnTU-sEzN-k6Kk-uoir86 encrypted 0 osd fsid 1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 osd id 0 osdspec affinity type db vdo 0 wal device /dev/vg-osd-wal/lv-osd-walc wal uuid JeSjju-3Lvc-3rCW-7dCH-44f3-2zAb-CPJyQG devices /dev/nvme1n1 [wal] /dev/vg-osd-wal/lv-osd-walc block device /dev/ceph-22fe0799-26b4-4e22-b28b-8eb1d4d0c42d/osd-block-1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 block uuid wMlS8p-nMeS-xRXm-R2fe-1kFM-MIYM-JWi2ie cephx lockbox secret cluster fsid ce5fe6a9-f807-4198-b8d4-c284322c82fa cluster name ceph crush device class None encrypted 0 osd fsid 1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 osd id 0 osdspec affinity type wal vdo 0 wal device /dev/vg-osd-wal/lv-osd-walc wal uuid JeSjju-3Lvc-3rCW-7dCH-44f3-2zAb-CPJyQG devices /dev/nvme0n1
'%3e%3c/path%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
'%3e%3c/path%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
我的osd服务节点上有10块7.3T的机械硬盘和3块7T的nvme盘。在我创建集群的时候，我把其中两块nvme盘分别做成了10个逻辑卷供osd做为wal和db使用。
我现在已经成功的创建了第一个osd，但是我开始怀疑它的合理性。假如我的一块nvme坏了的话，那么我这个节点上所有的osd都会失去wal或者db数据。ceph集群是否可以应对这种情况？或者我应该如何规划磁盘的使用才更加的合理？
====== osd.0 ======= [block] /dev/ceph-22fe0799-26b4-4e22-b28b-8eb1d4d0c42d/osd-block-1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 block device /dev/ceph-22fe0799-26b4-4e22-b28b-8eb1d4d0c42d/osd-block-1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 block uuid wMlS8p-nMeS-xRXm-R2fe-1kFM-MIYM-JWi2ie cephx lockbox secret cluster fsid ce5fe6a9-f807-4198-b8d4-c284322c82fa cluster name ceph crush device class None db device /dev/vg-osd-db/lv-osd-dbc db uuid OtPXxw-Oh9L-Mcur-tnTU-sEzN-k6Kk-uoir86 encrypted 0 osd fsid 1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 osd id 0 osdspec affinity type block vdo 0 wal device /dev/vg-osd-wal/lv-osd-walc wal uuid JeSjju-3Lvc-3rCW-7dCH-44f3-2zAb-CPJyQG devices /dev/sdc [db] /dev/vg-osd-db/lv-osd-dbc block device /dev/ceph-22fe0799-26b4-4e22-b28b-8eb1d4d0c42d/osd-block-1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 block uuid wMlS8p-nMeS-xRXm-R2fe-1kFM-MIYM-JWi2ie cephx lockbox secret cluster fsid ce5fe6a9-f807-4198-b8d4-c284322c82fa cluster name ceph crush device class None db device /dev/vg-osd-db/lv-osd-dbc db uuid OtPXxw-Oh9L-Mcur-tnTU-sEzN-k6Kk-uoir86 encrypted 0 osd fsid 1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 osd id 0 osdspec affinity type db vdo 0 wal device /dev/vg-osd-wal/lv-osd-walc wal uuid JeSjju-3Lvc-3rCW-7dCH-44f3-2zAb-CPJyQG devices /dev/nvme1n1 [wal] /dev/vg-osd-wal/lv-osd-walc block device /dev/ceph-22fe0799-26b4-4e22-b28b-8eb1d4d0c42d/osd-block-1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 block uuid wMlS8p-nMeS-xRXm-R2fe-1kFM-MIYM-JWi2ie cephx lockbox secret cluster fsid ce5fe6a9-f807-4198-b8d4-c284322c82fa cluster name ceph crush device class None encrypted 0 osd fsid 1c3acfe4-b3f8-4668-9a20-8ac98cf4f580 osd id 0 osdspec affinity type wal vdo 0 wal device /dev/vg-osd-wal/lv-osd-walc wal uuid JeSjju-3Lvc-3rCW-7dCH-44f3-2zAb-CPJyQG devices /dev/nvme0n1