I upgraded glusterfs to 3.3, created a new volume, then moved the data into the new volume. It took quite a long time, overall. After bringing data over, I started all instances on virt1, then on virt2, then on virt4, with a 30 second delay between instances. Then I selectively brought instances up on virt3 by priority order of the project. The only instances I didn't bring up on virt3 were from the following projects: juju, dumps, etherpad, puppet, testing, turnkey-mediawiki. I believe these projects are not currently in use.
If your instance wasn't brought up, or it failed to boot, you can start and reboot it via the reboot action in labsconsole. If you have rebooted your instance and are still having issues with it, please let me know. Hopefully gluster 3.3 solves some of our issues. 3.2.x was apparently ill-suited for hosting virtual machine images, where 3.3 has been designed to handle them well. If we have another major issue with gluster, I'll heavily consider moving to another solution. - Ryan _______________________________________________ Labs-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/labs-l
