[LU-5127] Running mke2fs to get alternate superblocks on >16TB LUN Created: 31/May/14 Updated: 03/Jun/14 Resolved: 03/Jun/14 |
|
| Status: | Closed |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.4.1 |
| Fix Version/s: | None |
| Type: | Question/Request | Priority: | Critical |
| Reporter: | Blake Caldwell | Assignee: | Jian Yu |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | None | ||
| Environment: |
RHEL 6.4 |
||
| Rank (Obsolete): | 14152 |
| Description |
|
We have an OST that is not mountable because a superblock is unreadable from the backend storage. In order to begin repairs, we need to run mke2fs to get the alternate superblocks to then run debugfs. Unfortunately, the current version of e2fsprogs does not handle this by default. Can you advise the correct options to use with mke2fs with -n (very important!). If we run with -O64bit without any other of the original filesystem options, will that do it? [root@f1-oss1d5 share]# /tmp/e2fsprogs-new/sbin/mke2fs -n /dev/mapper/f1-ddn1d-l30 [root@f1-oss1d5 man8]# dumpe2fs /dev/mapper/f1-ddn1d-l30 | grep superblock |
| Comments |
| Comment by Peter Jones [ 31/May/14 ] |
|
Yu, Jian Could you please advise here? Thanks Peter |
| Comment by Blake Caldwell [ 01/Jun/14 ] |
|
Hello, We were able to resolve this issue. The -t ext4 option was critical to get mke2fs to assume the 64bit feature. Here are the full commands that we used. [root@f1-oss1d1 ~]# mke2fs -n -t ext4 -O has_journal,ext_attr,resize_inode,dir_index,filetype,extent,64bit,mmp,flex_bg,sparse_super,large_file,huge_file,uninit_bg,dir_nlink,quota /dev/mapper/f1-ddn1d-l30 [root@f1-oss1d1 ~]# debugfs -b 4096 -s 32768 -c -R 'icheck 3992983288' /dev/mapper/f1-ddn1d-l30 |
| Comment by Jian Yu [ 03/Jun/14 ] |
|
Hi Blake, For -O option, you can only specify the features which are not listed in the [defaults] section and [fs_types] subsectionin in the /etc/mke2fs.conf configuration file. Is there anything else I can help in this ticket? |
| Comment by Blake Caldwell [ 03/Jun/14 ] |
|
That answers it. Thanks! |
| Comment by Jian Yu [ 03/Jun/14 ] |
|
You're welcome, Blake. May I close this ticket? |
| Comment by Blake Caldwell [ 03/Jun/14 ] |
|
Yes. Please do. |