Hide Forgot
Document URL: ============= https://access.redhat.com/documentation/en-US/Red_Hat_Storage/3.1/html/Installation_Guide/ Section Number and Name: ======================== 8.1. Updating Red Hat Gluster Storage from 3.0.x to 3.1.x in the Offline Mode 8.3. In-service Software Update from Red Hat Gluster Storage 3.y.z to Red Hat Gluster Storage 3.1.x Describe the issue: =================== gluster_shared_storage volume (or in other words, meta volume) if configured, needs to be remounted explicitly after upgrade. Starting glusterd will not take care of this. IF there is a reboot involved, then we will not end up in this situation as an entry in /etc/fstab will take care of auto-mounting it. However, we do not recommend a reboot if there is no kernel update. Then we land up in a situation where gluster_shared_storage volume will not be mounted in its standard path, resulting in some features which use meta-volume (geo-rep, snapshot scheduler, nfs-ganesha) not working correctly. Suggestions for improvement: ============================ In section 8.1, after step8 AND In section 8.3 after step10, Add an extra step mentioning: " IF there is no reboot involved, and if a meta-volume was configured before, then run the below command as is on ALL nodes of the cluster: mount /var/run/gluster/shared_storage/ IF the above command does not work, then something is wrong with the corresponding entry in /etc/fstab. Add the below line in /etc/fstab and run the command again. echo "<hostname/IP>:/gluster_shared_storage /var/run/gluster/shared_storage/ glusterfs defaults 0 0" >> /etc/fstab mount /var/run/gluster/shared_storage/ "
This is valid and needs to be added as mentioned in bug description.
Hi Anjana, Laura, This bug needs to be targeted for 3.2.0 and not for 3.2.0_async. This require an additional step to remount via (mount -a) in upgrade section. If we do not document this step, then there are high chances that customers from 3.1.3 to 3.2 will break if they dont upgrade the base kernel. We have been fortunate not to have a customer case around this so far mostly a reboot would have solved this. Please scope for 3.2.0. I will be ok to take this even post 3rd March.
That is correct, Laura.
Hi Laura, even though shared_storage and meta-volume mean the same and are used interchangeably, it would be necessary to stick to the word that gluster code understands . Thus, the command to be given instead of "mount metavolume" is "mount /var/run/gluster/shared_storage/" If the above command doesn't work, follow the steps below:. Step1: fix the corresponding entry in /etc/fstab "<Local Hostname/IP>:/gluster_shared_storage /var/run/gluster/shared_storage/ glusterfs defaults 0 0" Step2: Execute the mount command "mount /var/run/gluster/shared_storage/"
Reviewed the changes. Moving this BZ to verified in 3.2
Moving to CLOSED CURRENTRELEASE since RHGS 3.2 GA was yesterday. All documentation is available from https://access.redhat.com/documentation/en/red-hat-gluster-storage/.