Hide Forgot
Description of problem: On an already existing volume in a SSL enabled gluster cluster with samba ctdb. After creating a snapshot i am unable to see .snaps directory neither in windows client nor in a cifs mount. Where as everything works as expected when it comes to a fresh volume. After mounting a fresh volume and creating a snapshot the .snaps directory is seen. Version-Release number of selected component (if applicable): samba-client-libs-4.4.6-2.el7rhgs.x86_64 glusterfs-cli-3.8.4-7.el7rhgs.x86_64 WIndows10 How reproducible: Always with the existing volume Steps to Reproduce: 1.4 node cluster with samba ctdb & SSL enabled setup both for management and IO encryption 2.There must be an existing volume with USS enabled & vss plugin 3.create a snapshot & activate it 4.Mount the samba share in windows and check for the .snaps directory Actual results: No .snaps directory is available Expected results: .snaps directory should be visible Additional info: Volume Name: samba-official Type: Distributed-Replicate Volume ID: 8c090b71-a937-4869-9000-080cfdb40bc2 Status: Started Snapshot Count: 0 Number of Bricks: 2 x 2 = 4 Transport-type: tcp Bricks: Brick1: dhcp43-214.lab.eng.blr.redhat.com:/bricks/brick0/samba-official_brick0 Brick2: dhcp42-236.lab.eng.blr.redhat.com:/bricks/brick0/samba-official_brick1 Brick3: dhcp43-77.lab.eng.blr.redhat.com:/bricks/brick0/samba-official_brick2 Brick4: dhcp41-169.lab.eng.blr.redhat.com:/bricks/brick0/samba-official_brick3 Options Reconfigured: performance.readdir-ahead: on server.allow-insecure: on performance.stat-prefetch: on storage.batch-fsync-delay-usec: 0 features.cache-invalidation: on features.cache-invalidation-timeout: 600 performance.cache-samba-metadata: on performance.cache-invalidation: on performance.md-cache-timeout: 600 features.quota: on features.inode-quota: on client.ssl: on server.ssl: on features.quota-deem-statfs: on features.uss: enable features.show-snapshot-directory: enable features.barrier: disable features.soft-timeout: 0s features.hard-timeout: 0s
Can you be little more clear on the steps to reproduce the issue? What is the difference between the new volume and the existing volume? Did you first enabled SSL and then USS or the other way around? Can you give clear set of instructions to reproduce the issue in my setup?
The volume in question was created on an earlier gluster version i.e glusterfs-cli-3.8.4-6.el7rhgs.x86_64. First i enabled the SSL and then USS. Did some basic operations over that volume. Then i went on to upgrade my setup from 3.8.4-6 to 3.8.4-7. And then when i tried to create snapshot and activated it i am not seeing .snaps directory. I created a fresh volume over the 3.8.4-7 build and enabled USS. I took snapshot and was able to see .snap directory. **The issue is with that particular volume only which was present from the earlier build. I will follow the same steps and try to reproduce it again i.e a)Have a setup with previous build or setup with gluster 3.8.4-6 b)Create a volume and enable SSl c)Enable uss with all the VSS plugins d)Perform random operations along with snapshot test cases e)Upgrade the setup with the latest build f)Now try to take snapshot and see for .snaps folder availability.
This seems like the IPV6 issue. Can you please recheck this? Otherwise my suggestion is to close this bug.
Putting a need info on qe based on comment 7
I worked on the steps to reproduce but I am unable to reproduce this issue again.
Vivek, Can you please close the bz based on comment7 and comment10. Regards Rafi KC