Red Hat Bugzilla – Bug 829808
Can't create a volume with the same remote directories after deleting one
Last modified: 2015-11-03 18:04:37 EST
Description of problem:
When deleting a volume using the CLI, you can't directly re-create a volume with the same parameters of remote partitions to use.
Getting an error saying that the volume is already used
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Create a volume on a blank install of Gluster FS : sudo gluster volume create toto xxx.xxx.xxx.xxx:/titi xxx.xxx.xxx.xxx:/titi
2. Delete the volume : sudo gluster volume delete toto (yes)
3. Verify that the volume has been destroyed : sudo gluster volume info
4. Create the same volume sudo gluster volume create toto xxx.xxx.xxx.xxx:/titi xxx.xxx.xxx.xxx:/titi
Error : /titi or a prefix name is already part of a volume
Volume correctly re-created.
After deep analysis and quite lots of time, some attributs were still defined on /titi folder on all machines.
Seen using the following command on each machines :
sudo getfattr --absolute-names -m . -d -e hex /titi
Corrected using the following command with all attributs found :
sudo setfattr -x "xxx" /titi
Looks like the CLI isn't deleting the volume correctly...
*** Bug 829811 has been marked as a duplicate of this bug. ***
This is currently intended behaviour. See http://review.gluster.com/781
As this 'problem' seems to happen a lot, it might be a good idea to have the xattrs removed when a brick is removed from a volume, or the volume is deleted.
Actually, I didn't even heard about this script on the Net while searching for a solution, it could be documented somewhere or highlighted in the manual, but it could be pretty interesting to be a native behaviour.
If you want to keep this feature apart, could you please document it in the manual in order to users to know that they wont be able to re-create a volume afterward, unless they run the script.
Thanks for your fast answer
*** This bug has been marked as a duplicate of bug 812214 ***