lustre 文件系统无法格式化 10T磁盘分区 求解决

bazhic

[root@MDS ~]# mkfs.lustre --reformat --backfstype=ldiskfs --fsname=test2 --mdt --mgs /dev/sdb3

   Permanent disk data:
Target:     test2-MDTffff
Index:      unassigned
Lustre FS:  test2
Mount type: ldiskfs
Flags:      0x75
              (MDT MGS needs_index first_time update )
Persistent mount opts: iopen_nopriv,user_xattr,errors=remount-ro
Parameters: mdt.group_upcall=/usr/sbin/l_getgroups

device size = 9536743MB
2 6 18
formatting backing filesystem ldiskfs on /dev/sdb3
        target name  test2-MDTffff
        4k blocks     2441406250
        options        -J size=400 -i 4096 -I 512 -q -O dir_index,uninit_groups -F
mkfs_cmd = mke2fs -j -b 4096 -L test2-MDTffff  -J size=400 -i 4096 -I 512 -q -O dir_index,uninit_groups -F /dev/sdb3 2441406250
mkfs.lustre: Unable to mount /dev/sdb3: Invalid argument

mkfs.lustre FATAL: failed to write local files
mkfs.lustre: exiting with 22 (Invalid argument)

bazhic

无法格式化分区/dev/sdb3  这个分区大小是10T   回复 2# Hongqiyaodao


   

bazhic

不是应该先格式化 后才能挂载吗？  6T的分区 就没有问题  我应该如果解决。回复 4# Hongqiyaodao


   

chenyx

是不是你的那个分区格式有容量的限制,你搜索下资料,看看你的那个lustre格式的最大容量支持是多少

bazhic

我也没找到  我这有一些帮助信息 ，请你帮助我一起看看呗  

[root@MDS ~]# mkfs.lustre --help
mkfs.lustre v1.8.7
usage: mkfs.lustre <target types> [options] <device>
        <device>:block device or file (e.g /dev/sda or /tmp/ost1)
        target types:
                --ost: object storage, mutually exclusive with mdt,mgs
                --mdt: metadata storage, mutually exclusive with ost
                --mgs: configuration management service - one per site
        options (in order of popularity):
                --mgsnode=<nid>[,<...>] : NID(s) of a remote mgs node
                        required for all targets other than the mgs node
                --fsname=<filesystem_name> : default is 'lustre'
                --failnode=<nid>[,<...>] : NID(s) of a failover partner
                        cannot be used with --servicenode
                --servicenode=<nid>[,<...>] : NID(s) of all service partners
                        treat all nodes as equal service node, cannot be used with --failnode
                --param <key>=<value> : set a permanent parameter
                        e.g. --param sys.timeout=40
                             --param lov.stripesize=2M
                --index=#N : target index (i.e. ost index within the lov)
                --comment=<user comment>: arbitrary user string (824 bytes)
                --mountfsoptions=<opts> : permanent mount options
                --backfstype=<fstype> : backing fs type (ext3, ldiskfs)
                --device-size=#N(KB) : device size for loop devices
                --mkfsoptions=<opts> : format options
                --reformat: overwrite an existing disk
                --stripe-count-hint=#N : used for optimizing MDT inode size
                --dryrun: just report what we would do; don't write to disk
                --verbose : e.g. show mkfs progress
                --quiet回复 7# Hongqiyaodao


   

zongg

估计有限制,没做过这么大的.

bazhic

这是lustre的介绍
Lustre System Capacity
Lustre file system capacity is the sum of the capacities provided by the targets.
As an example, 64 OSSs, each with two 8-TB targets, provide a file system with a
capacity of nearly 1 PB. If this system uses sixteen 1-TB SATA disks, it may be
possible to get 50 MB/sec from each drive, providing up to 800 MB/sec of disk
bandwidth. If this system is used as storage backend with a system network like
InfiniBand that supports a similar bandwidth, then each OSS could provide 800
MB/sec of end-to-end I/O throughput. Note that the OSS must provide inbound and
outbound bus throughput of 800 MB/sec simultaneously. The cluster could see
aggregate I/O bandwidth of 64x800, or about 50 GB/sec. Although the architectural
constraints described here are simple, in practice it takes careful hardware selection,
benchmarking and integration to obtain such results.
In a Lustre file system, storage is only attached to server nodes, not to client nodes. If
failover capability is desired, then this storage must be attached to multiple servers.
In all cases, the use of storage area networks (SANs) with expensive switches can be
avoided, because point-to-point connections between the servers and the storage
arrays normally provide the simplest and best attachments.回复 9# zongg