[Bugs] [Bug 1151004] New: [USS]: deletion and creation of snapshots with same name causes problems

bugzilla at redhat.com bugzilla at redhat.com
Thu Oct 9 11:56:35 UTC 2014


https://bugzilla.redhat.com/show_bug.cgi?id=1151004

            Bug ID: 1151004
           Summary: [USS]: deletion and creation of snapshots with same
                    name causes problems
           Product: GlusterFS
           Version: mainline
         Component: snapshot
          Assignee: gluster-bugs at redhat.com
          Reporter: rabhat at redhat.com
                CC: bugs at gluster.org



Description of problem:

The deletion and recreation of a snapshot with same name creates problems while
accessing the contents of newly created snapshot bu giving ENOTCONN errors.


Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
1. Create a file on the glusterfs mount point
2. Create a snapshot "snap1"
3. delete the file
4. delete the snapshot "snap1"
5. create a snapshot "snap1"
6. try to access snap1

Actual results:
applications get ENOTCONN

Expected results:
Applications should not get ENOTCONN

Additional info:
When the snapshot "snap1" is created first time and accessed (<mount
point>/<entry point>/snap1/<filename>), a new inode and dentry were created for
snap1 with parent being the entry point. Within the inode context the glfs_t
instance to access the snapshot and the handle to access the inode in the
snapshot world (or gfapi world in this case) is saved. But when the snapshot
was deleted and recreated, a new glfs_t instance is established (older one
being destroyed). So to access snap1 now, we have to do yet another lookup and
save the glfs_t instance and the new handle within the inode context. Since
that was not being done, older glfs_t instance was still used which gave
ENOTCONN when accessed as the corresponding snapshot volume does not exist
anymore.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=hmPfgd1kd4&a=cc_unsubscribe


More information about the Bugs mailing list