[Ocfs2-devel] File system is read-only review

Wed Mar 23 23:11:13 PDT 2016

On 2016/3/24 13:49, Gang He wrote:
> Hi Jeseph and Jiufei,
> 
> 
>>>>
>> Hi Gang，
>>
>> On 2016/3/23 17:46, Gang He wrote:
>>> Hello Joseph and Jiufei,
>>>
>>> I got a customer issue, ocfs2: filesystem goes read-only after node restart.
>>> The symptom is very similar with this bug, when the ocfs2 file system called 
>> ocfs2_recover_orphans() during mounting period, 
>>> Then, encountered this issue, the stack likes,
>>> Mar 10 17:45:11 mipkhpch2bl07 kernel: [   81.689531] ocfs2: Mounting device 
>> (253,1) on (node 1770803, slot 0) with writeback data mode.
>>> Mar 10 17:45:11 mipkhpch2bl07 kernel: [   81.695279] OCFS2: ERROR (device 
>> dm-1): ocfs2_commit_truncate: Inode 1075718 has empty extent block at 1075718
>>> Mar 10 17:45:11 mipkhpch2bl07 kernel: [   81.695283]
>>> Mar 10 17:45:11 mipkhpch2bl07 kernel: [   81.695287] File system is now 
>> read-only due to the potential of on-disk corruption. Please run fsck.ocfs2 
>> once the file system is unmounted.
>>> Mar 10 17:45:11 mipkhpch2bl07 kernel: [   81.695297] 
>> (kworker/u:36,913,16):ocfs2_truncate_for_delete:585 ERROR: status = -30
>>> Mar 10 17:45:11 mipkhpch2bl07 kernel: [   81.695302] 
>> (kworker/u:36,913,16):ocfs2_wipe_inode:754 ERROR: status = -30
>>> Mar 10 17:45:11 mipkhpch2bl07 kernel: [   81.695308] 
>> (kworker/u:36,913,16):ocfs2_delete_inode:1039 ERROR: status = -30
>>>
>>> I saw your mails, Jiufei ever tried to fix this bug, but these three patches 
>> were not all accepted by Linux git tree.
>>> Then, I want to know what is this bug's final status? 
>> OCFS2 readonly due to corrupted extent block or extent record.
>>
>>> Second, as you said, misusing jbd2_journal_restart() possibly results in 
>> this issue, by now, which patch can fix this misusing?
>> The patch series submitted by Jiufei.
>> https://oss.oracle.com/pipermail/ocfs2-devel/2015-April/010768.html 
>> There are still two patches are pending in the next tree:
>> ocfs2: extend enough credits for freeing one truncate record while replaying 
>> truncate records
>> ocfs2: extend transaction for ocfs2_remove_rightmost_path() and 
>> ocfs2_update_edge_lengths() before to avoid inconsistency between inode and 
>> et
> Thank for your comments, one more question.
> Only merging these three patches can fix this bug? my means is , are there any other dependent patch need to be merge firstly?
> since we saw Jiufei submited a few patches continuously in this area in the past year.
Yes, only these three patches are jbd2_journal_restart() related.

Thanks,
Joseph

> 
> Thanks a lot.
> Gang 
> 
>>
>>> Finally, if the customers have hit this case (very bad, start ocfs2 file 
>> system and get read-only), 
>>> how to get the file system back? unmount the file system and run fsck.ocfs2, 
>> or patch the kernel code and bypass this logic(which patch can do this?). 
>> Merge these patches can only correct the logic in the future.
>> For the corrupted filesystem, you should run fsck.
>>
>> Thanks,
>> Joseph
>>
>>>  
>>>
>>> Thanks
>>> Gang 
>>>
>>>
>>>
>>>>>>
>>>> Hi Guanghui,
>>>> Misusing jbd2_journal_restart() may cause this issue.
>>>> You can find details and patches we sent before.
>>>>
>>>> ocfs2: fix some inconsistency problems caused by jbd2_journal_restart()
>>>> https://oss.oracle.com/pipermail/ocfs2-devel/2015-April/010768.html 
>>>>
>>>> On 2015/11/20 19:00, Zhangguanghui wrote:
>>>>>
>>>>> File system is read-only when deleting the inode, So I try fsck.ocfs2 -f and 
>>
>>>> success,
>>>>> Anybody ever see this, how to avoid this, Thanks 
>>>>>
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.362978] OCFS2: ERROR (device dm-2): 
>>>> ocfs2_commit_truncate: Inode 24772873 has empty extent block at 24772873
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.362978]
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.362986] File system is now read-only 
>>>> due to the potential of on-disk corruption. Please run fsck.ocfs2 once the 
>>>> file system is unmounted.
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.362991] 
>>>> (kworker/u129:0,32928,0):ocfs2_truncate_for_delete:574 ERROR: status = -30
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.362993] 
>>>> (kworker/u129:0,32928,0):ocfs2_wipe_inode:743 ERROR: status = -30
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.362996] 
>>>> (kworker/u129:0,32928,0):ocfs2_delete_inode:1022 ERROR: status = -30
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.434249] OCFS2: ERROR (device dm-2): 
>>>> ocfs2_commit_truncate: Inode 1445125 has empty extent block at 1445125
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.434249]
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.434256] 
>>>> (kworker/u129:0,32928,1):ocfs2_truncate_for_delete:574 ERROR: status = -30
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.434258] 
>>>> (kworker/u129:0,32928,1):ocfs2_wipe_inode:743 ERROR: status = -30
>>>>> Nov 18 11:04:30 cvk110 kernel: [144411.434261] 
>>>> (kworker/u129:0,32928,1):ocfs2_delete_inode:1022 ERROR: status = -30
>>>>>
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>>
>> -----------------------------------------------------------------------------
>>>> ---------------------------------------------------------------!
>>>> ---
>>>>> zhangguanghui
>>>>>
>>>>
>> -----------------------------------------------------------------------------
>>>> --------------------------------------------------------
>>>>> 本邮件及其附件含有杭州华三通信技术有限公司的保密信息仅限于发送给上面地址中列出
>>>>> 的个人或群组。禁止任何其他人以任何形式使用（包括但不限于全部或部分地泄露、复制、
>>>>> 或散发）本邮件中的信息。如果您错收了本邮件，请您立即电话或邮件通知发件人并删除本
>>>>> 邮件！
>>>>> This e-mail and its attachments contain confidential information from H3C, 
>>>> which is
>>>>> intended only for the person or entity whose address is listed above. Any 
>>>> use of the
>>>>> information contained herein in any way (including, but not limited to, 
>>>> total or partial
>>>>> disclosure, reproduction, or dissemination) by persons other than the 
>>>> intended
>>>>> recipient(s) is prohibited. If you receive this e-mail in error, please 
>>>> notify the sender
>>>>> by phone or email immediately and delete it!
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Ocfs2-devel mailing list
>>>>> Ocfs2-devel at oss.oracle.com 
>>>>> https://oss.oracle.com/mailman/listinfo/ocfs2-devel 
>>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Ocfs2-devel mailing list
>>>> Ocfs2-devel at oss.oracle.com 
>>>> https://oss.oracle.com/mailman/listinfo/ocfs2-devel 
>>>
>>> .
>>>
> 
> .
>