Red Hat Bugzilla – Bug 851962
Geo-rep session started successfully even though one of the brick machine didn't have info about the session.
Last modified: 2014-08-24 20:49:43 EDT
Description of problem:Geo-rep session started successfully, even though it couldn't write the info about the geo-rep session in the file /var/lib/glusterd/vols/master/info in one of the brick machines. The reason it couldn't write the info was , that machine didn't have space in the root partition.
Version-Release number of selected component (if applicable): RHS-2.0.z
How reproducible: May be not consistent.
Steps to Reproduce:
1.Create dist-rep volume for master.
2.Fill up the root partition of one of the brick machine of the master.
3.Try to start geo-rep session.
Actual results: the geo-rep start command succeeded
Expected results: It should fail with proper error message.
Could you please attache the logs if possible?
fixed with newer geo-replication implementation.
since geo-rep got revamped since this bug, now the behavior is, if it fails to write to the file, it fails at commit stage, it doesn't fail completely, in the sense it will have inconsistent data over the cluster. There is a bug for that
verified on glusterfs-18.104.22.168rhs-1.el6rhs.x86_64
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA.
For information on the advisory, and where to find the updated files, follow the link below.
If the solution does not work for you, open a new bug report.