[Ocfs2-devel] Different idea about slot overwritten

[Guozhonghua]

Hi Jiang,



I think there is another scenario about slot overwritten issue.

There are three nodes in the ocfs2 cluster. Node 3 had mounted with slot 1.

Node 1 and node 2 execute mounting volume operation at the same time.



        N1                               N2

        mount ocfs2 volume               mount ocfs2 volume

        ocfs2_fill_super()                 ocfs2_fill_super()

          ocfs2_initialize_super            ocfs2_initialize_super

            ... ...                          ... ...

            ocfs2_init_slot_info(osb);       ocfs2_init_slot_info(osb);

      ocfs2_mount_volume               ocfs2_mount_volume

        ocfs2_super_lock                 ocfs2_super_lock

         Gotten the super lock             Waiting for the super lock

         Find slot 0 unused

         from memory

                update the slot 0 with 1

                ... ...

                 locked journal 0

        mount finished.

                                              Gotten super lock and

                                                                                  Also find slot 0 unused

                                                                                  from memory,

                                                                                  update the slot 0 with node num 2

                                                                                  But Journal 0 is locked by N1

                                                                                  Mounted hang up.

          ... ...                             ... ...

    umount volume                         ... ...

         cleare the slot 0                ... ...

                                          Gotten joural 0 lock

                                     mount finished.

                                                                         But here, the slot 0 is cleare by N1



        IF N1 mount again

        Same condition with N2

        and will hang up.



In the function of ocfs2_mount_volume, I think the slot info should be refreshed after ocfs2_super_lock called.

static int ocfs2_mount_volume(struct super_block *sb)

{

        status = ocfs2_super_lock(osb, 1);

        ......



+      status = ocfs2_refresh_slot_info(osb);

+      if (status < 0) {

+              mlog_errno(status);

+              goto leave;

+      }

        ... ...

}



Another way is to move ocfs2_init_slot_info() function from ocfs2_initialize_super to replace ocfs2_refresh_slot_info as above.





Message: 5

Date: Wed, 23 Dec 2015 18:23:36 +0800

From: jiangyiwen <jiangyiwen at huawei.com<mailto:jiangyiwen at huawei.com>>

Subject: [Ocfs2-devel] [PATCH] ocfs2: fix slot overwritten if storage

        link  down during mount

To: Andrew Morton <akpm at linux-foundation.org<mailto:akpm at linux-foundation.org>>

Cc: Mark Fasheh <mfasheh at suse.de<mailto:mfasheh at suse.de>>, ocfs2-devel at oss.oracle.com<mailto:ocfs2-devel at oss.oracle.com>

Message-ID: <567A7628.5040503 at huawei.com<mailto:567A7628.5040503 at huawei.com>>

Content-Type: text/plain; charset="utf-8"



The following case will lead to slot overwritten.



N1                               N2

mount ocfs2 volume, find and

allocate slot 0, then set

osb->slot_num to 0, begin to

write slot info to disk

                                 mount ocfs2 volume, wait for super lock

write block fail because of

storage link down, unlock

super lock

                                 got super lock and also allocate slot 0

                                 then unlock super lock



mount fail and then dismount,

since osb->slot_num is 0, try to

put invalid slot to disk. And it

will succeed if storage link

restores.

                                 N2 slot info is now overwritten



-------------------------------------------------------------------------------------------------------------------------------------
????????????????????????????????????????
????????????????????????????????????????
????????????????????????????????????????
???
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!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://oss.oracle.com/pipermail/ocfs2-devel/attachments/20151225/c5b6998e/attachment.html