[Ocfs2-users] Read only issue, any other ideas aout it will be appreciated, Thanks a lot.

[Goldwyn Rodrigues]

Hi Guozhonghua,

On 02/13/2015 02:39 AM, Guozhonghua wrote:
> Hi All,
>
> As we test with OCFS2, sometime the file system become read only. The
> log is bellows.
>
> We had review the code and try to find the reason, but did not find any
> ways to avoid this issue.
>
> All the nodes in the cluster reach the read only file system finally.
>
> Feb 12 15:38:30 node01 kernel: [173109.179478] OCFS2: ERROR (device
> dm-1): ocfs2_commit_truncate: Inode 8918280 has empty extent block at
> 8918280
>
> Feb 12 15:38:30 node01 kernel: [173109.179478]
>
> Feb 12 15:38:30 node01 kernel: [173109.179493] File system is now
> read-only due to the potential of on-disk corruption. Please run
> fsck.ocfs2 once the file system is unmounted.
>
> Feb 12 15:38:30 node01 kernel: [173109.232387]
> (libvirtd,7367,11):ocfs2_truncate_for_delete:574 ERROR: status = -30
>
> Feb 12 15:38:30 node01 kernel: [173109.232401]
> (libvirtd,7367,11):ocfs2_wipe_inode:743 ERROR: status = -30
>
> Feb 12 15:38:30 node01 kernel: [173109.232408]
> (libvirtd,7367,11):ocfs2_delete_inode:1022 ERROR: status = -30
>
> We had use the command of “fsck.ocfs –fpy” to repair the file system.
>
> The repair procedures log is as below, there are some error about the
> inode data:
>
> fsck.ocfs2 1.6.3
>
> Checking OCFS2 filesystem in /dev/dm-1:
>
>    Label:              o20140820105531
>
>    UUID:               F1AAAC6C765D480F9052946392C6527A
>
>    Number of blocks:   2621440000
>
>    Block size:         4096
>
>    Number of clusters: 10240000
>
>    Cluster size:       1048576
>
>    Number of slots:    32
>
> /dev/dm-1 was run with -f, check forced.
>
> Pass 0a: Checking cluster allocation chains
>
> [CHAIN_BITS] Chain 60 in allocator inode 519 has 36784 bits marked free
> out of 64512 total bits but the block groups in the chain have 36762
> free out of 64512 total.  Fix this by updating the chain record? y
>
> [CHAIN_GROUP_BITS] Allocator inode 519 has 2071620 bits marked used out
> of 10240000 total bits but the chains have 2071642 used out of 10240000
> total.  Fix this by updating the inode counts? y
>
> Pass 0b: Checking inode allocation chains
>
> Pass 0c: Checking extent block allocation chains
>
> ^[Pass 1: Checking inodes and blocks.
>
> [INODE_SPARSE_CLUSTERS] Inode 8918280 has 4294967286 clusters but its
> blocks fit in 0 clusters. Correct the number of clusters? y
>
> Pass 2: Checking directory entries.
>
> Pass 3: Checking directory connectivity.
>
> Pass 4a: checking for orphaned inodes
>
> [INODE_ORPHANED] Inode 8918280 was found in the orphan directory. Delete
> its contents and unlink it? y
>
> Pass 4b: Checking inodes link counts.
>
> [INODE_NOT_CONNECTED] Inode 8918281 isn't referenced by any directory
> entries.  Move it to lost+found? y
>
> [INODE_NOT_CONNECTED] Inode 8918282 isn't referenced by any directory
> entries.  Move it to lost+found? y
>
> All passes succeeded.
>
> We are using Linux kernel with version 3.13.
>
> Is there any way to fix it? Or any other patches?
>

This does look problematic if this is happening again and again. Does it 
happen on the same file? Did you try and figure out what the inode 
8918280 is, and what kind of access pattern is done on that file?

-- 
Goldwyn