Description of problem: ----------------------- While performing upgrade, the check is done whether distributed volume is UP. If there are distributed volume running, then the upgrading glusterfs package fails. But now, glusterfs-client-xlators package is getting installed, even when the distributed volume is up Version-Release number of selected component (if applicable): ------------------------------------------------------------- glusterfs-3.7.1-3.el6rhs How reproducible: ----------------- Always/consistent Steps to Reproduce: ------------------- 1. Create and start the distribute volume with RHGS 3.0.4 2. Try to update gluster rpms to RHGS 3.1 ( glusterfs-3.7.1-3.el6rhs ) Actual results: --------------- glusterfs-client-xlators package is getting installed, while other gluster rpms are abandoned to install Expected results: ----------------- gluster core packages should not get installed or upgraded when distributed volume is up
There was a similar bug with RHGS 3.0.4,where glusterfs-geo-replication and glusterfs-cli packages are getting updated, when distributed volume is up. This issue also had a customer case attached to it. This issue is tracked in https://bugzilla.redhat.com/show_bug.cgi?id=1211839 and was resolved with the latest glusterfs build ( glusterfs-3.7.1-3.el6rhs )
As per Bala's email there are 3 packages that are getting affected - When performing in-service software update, when distributed volume is up and those are, 1. glusterfs-ganesha 2. glusterfs-client-xlators 3. python-gluster Changing the bug summary accordingly
Adding the comment from Niels on that mail thread <snip> Now, all sub-packages of the glusterfs src.rpm will need the %pretrans scripts. If a package does not have the script, it might get updated while Gluster processes are running. This is not a problem, until the processes get restarted and different versions of libraries are expected. Likely no immediate errors, but hard to debug unexpected behaviour could be the result. Maybe the %pretrans is not needed for python-gluster, but it should be required for the others. Any sub-package that has a versioned dependency on any of the glusterfs packages needs the %pretrans script. </snip>
Patch is under review at https://code.engineering.redhat.com/gerrit/50967
Tested with glusterfs-3.7.1-4.el6rhs, Tried to update glusterfs rpms ( from RHGS 3.0.4 to RHGS 3.1 ), with distributed volume as well as without stopping the brick process. On both this occasion, gluster core packages ( glusterfs-*, glusterfs-client-xlators, glusterfs-ganesha ) are not installed/updated. Marking this bug as VERIFIED
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. https://rhn.redhat.com/errata/RHSA-2015-1495.html