------------------------------------------------------------------- Wed May 06 07:35:44 UTC 2015 - vuntz@suse.com - Track suse-cloud-5 git branch. - Update to version 5+git.1430897639.60f8b6d: + Fix attribute path for repositories in provisioner.json + Fix attribute path for repositories in provisioner.json in pre ------------------------------------------------------------------- Tue Mar 31 07:41:59 UTC 2015 - vuntz@suse.com - Update to version 5+git.1427787684.4532af7: + Remove ceph package on upgrade from Cloud 4 ------------------------------------------------------------------- Thu Mar 05 10:05:32 UTC 2015 - vuntz@suse.com - Update to version 5+git.1425548496.4175cc2: + Fix update of repos in provisioner proposal (bnc#919957) + Test saving the provisioner proposal ------------------------------------------------------------------- Thu Mar 5 02:38:54 UTC 2015 - cloud-devel@suse.de - Update to version 5+git.1425465604.dbeebf4: + Correct instructions on how to update admin installation + Do not accept starting the upgrade if some nodes are not ready + Remove old crowbar db migration script breaking upgrade ------------------------------------------------------------------- Fri Feb 27 14:59:28 UTC 2015 - vuntz@opensuse.org - Update to version 5+git.1424873281.8f832f4: + Add comment to note a change we can do in the next version + Migrate pacemaker proposals to use udpu + Do not fail if there's no neutron-ns-metadata-proxy process on a L3 node + Remove check that tells us that neutron-ns-metadata-proxy is running - Drop pull-request-45.patch: merged upstream. - Drop pull-request-46.patch: merged upstream. - Drop pull-request-47.patch: merged upstream. ------------------------------------------------------------------- Thu Feb 19 15:40:43 UTC 2015 - vuntz@suse.com - Add pull-request-47.patch: Remove check that tells us that neutron-ns-metadata-proxy is running ------------------------------------------------------------------- Thu Feb 19 09:32:03 UTC 2015 - vuntz@suse.com - Add pull-request-46.patch: Do not fail if there's no neutron-ns-metadata-proxy process on a L3 node ------------------------------------------------------------------- Wed Feb 18 22:19:43 UTC 2015 - cloud-devel@suse.de - Update to version 5+git.1424271951.9c50c46: + SUSE-Cloud-5-Pool repo now exists and can be checked + Check that SUSE-Cloud-5-Updates is correctly signed + Check that Cloud ISO is correctly signed ------------------------------------------------------------------- Wed Feb 18 22:13:52 UTC 2015 - vuntz@suse.com - Add pull-request-45.patch: Migrate pacemaker proposals to use udpu ------------------------------------------------------------------- Wed Feb 18 08:02:54 UTC 2015 - cloud-devel@suse.de - Update to version 5+git.1424183539.65ed5a3: + Complete fix for upgrade when some barclamps are not installed ------------------------------------------------------------------- Fri Feb 13 02:58:11 UTC 2015 - cloud-devel@suse.de - Update to version 5+git.1423750662.dd18131: + Do not stop group resources when stopping HA resources + Fix error message to mention "admin server" instead of "admin node" ------------------------------------------------------------------- Wed Feb 11 10:04:24 UTC 2015 - vuntz@suse.com - Update to version 5+git.1423648465.22d52c8: + Ignore when barclamp-tempest is missing. + Added checks for presence of Cloud5 repositories. + Move fetching the list of neutron-l3 nodes a bit later + Also move checks requiring access to nodes a bit later ------------------------------------------------------------------- Mon Feb 09 10:24:45 UTC 2015 - opensuse-packaging@opensuse.org - Update to version 5+git.1423243478.0c050ea: + Set frame_max setting of rabbitmq to 0 during upgrade (bnc#910815) + Do not check for status of /etc/init.d/openstack-neutron-ovs-cleanup - remove pull-request-37.patch, pull-request-38.patch: merged upstream ------------------------------------------------------------------- Thu Feb 5 18:01:44 UTC 2015 - vuntz@suse.com - Add pull-request-38.patch: Do not check for status of /etc/init.d/openstack-neutron-ovs-cleanup ------------------------------------------------------------------- Wed Feb 4 16:42:57 UTC 2015 - vuntz@suse.com - Add pull-request-37.patch: Set frame_max setting of rabbitmq to 0 during upgrade (bnc#910815) ------------------------------------------------------------------- Wed Feb 04 08:27:44 UTC 2015 - vuntz@suse.com - Update to version 5+git.1423038450.a46f16a: + Import repo directories from old Cloud 4 structure + Also create symlinks to the SLE12 mirrored repos + Re-sign PTF repos at the end of the upgrade process ------------------------------------------------------------------- Sat Jan 31 02:51:57 UTC 2015 - cloud-devel@suse.de - Update to version 5+git.1422631251.3f1ff30: + Manually kill neutron-ns-metadata-proxy + Make sure all openstack services are down on all nodes before we stop rabbitmq + only print warning about neutron-proxy process ------------------------------------------------------------------- Tue Jan 20 02:53:28 UTC 2015 - cloud-devel@suse.de - Update to version 5+git.1421694566.89b7e2e: + Stop and start chef services that are impacted by ruby version change + Remove the cached crowbar.yml files in -post before barclamp_install.rb + Generate new /opt/dell/crowbar_framework/config/secrets.env file in post + Add a check for barclamps not shipped by SUSE Cloud + Create /etc/chef/amqp_passwd + Fix re-installing optional barclamps + Update /etc/chef/{server,solr.rb} to read amqp password from file + Also stop and start couchdb and rabbitmq + Also stop epmd after we stopped rabbitmq + Also kill /opt/dell/bin/looper_chef_client.sh in pre + Prepare the Ha-enabled system for the upgrade: For each cluster (defined by cluster founder node), stop all pacemaker resources before the upgrade (before the openstack services are stopped). ------------------------------------------------------------------- Thu Dec 4 13:39:29 UTC 2014 - dmueller@suse.com - Update to version 5+git.1417593656.f40f142: + Fixed wrong update checking + Remove no longer needed neutron helper + Fix wrong directory name + Also deactivate trove and tempest - removed pull-request-22.patch: merged upstream ------------------------------------------------------------------- Fri Oct 10 08:22:18 UTC 2014 - dmueller@suse.com - Update to version 5+git.1412929332.7de8163: + Fix barclamp install order ------------------------------------------------------------------- Fri Oct 10 08:18:28 UTC 2014 - dmueller@suse.com - Update to version 5+git.1412070754.d66d26b: + Fix log in get_allocated_nodes to mention role + Make get_allocated_nodes always set get_allocated_nodes_retval as retval + Use local variables in functions + fix post step error message + Also disable openvswitch-switch ------------------------------------------------------------------- Fri Oct 10 08:17:32 UTC 2014 - dmueller@suse.com - add pull-request-22.patch ------------------------------------------------------------------- Mon Sep 22 11:53:03 UTC 2014 - dmueller@suse.com - Update to version 5+git.1411054777.a288e12: + fix post step error message + fix test for user answering Y + Prepare for Cloud 4->5 + Use xpath command instead of ugly xmllint command + Fix crash in post when /etc/crowbar/provisioner.json exists with no repo + Check the nodes have packages updated to SUSE Cloud 4 packages ------------------------------------------------------------------- Tue Sep 2 20:29:04 UTC 2014 - dmueller@suse.com - bump version to 5 ------------------------------------------------------------------- Tue Aug 26 12:00:26 UTC 2014 - vuntz@suse.com - Add pull-request-9.patch: Fix crash in post when /etc/crowbar/provisioner.json exists with no repo ------------------------------------------------------------------- Wed Aug 20 10:31:43 UTC 2014 - vuntz@suse.com - Update to version 4+git.1407997338.74f1df1: + We need to find the neutron-server nodes before the roles are removed - Drop bnc891545-neutron-server-detection.patch: merged upstream. - Add pull-request-5.patch: Check the nodes have packages updated to SUSE Cloud 4 packages (bnc#892500). - Add pull-request-6.patch: Remove duplicated agents in neutron database before doing db migration (bnc#892503). - Add pull-request-7.patch: Make get_allocated_nodes always set get_allocated_nodes_retval as retval ------------------------------------------------------------------- Tue Aug 12 12:12:03 UTC 2014 - nkrinner@suse.com - bnc#891545 - Upgrade scripts: neutron-server detection is broken Searching neutron-server nodes now happens when the relevant roles are still applied to the nodes ------------------------------------------------------------------- Fri Aug 1 07:32:46 UTC 2014 - vuntz@suse.com - Update to version 4+git.1406878311.148d161: + Adjust scripts to upgrade to Cloud 4 + Initial pass at upgrade support from Cloud 3 to Cloud 4 ------------------------------------------------------------------- Tue Jul 8 11:57:42 UTC 2014 - vuntz@suse.com - Bump version to 4 (no code change yet). ------------------------------------------------------------------- Tue Jan 28 13:39:13 UTC 2014 - vuntz@suse.com - Update to version 3+git.1390916278.8c83d30: + Add a helper to fix neutron database after upgrade + Add a FIXME for non-critical issue + Redownload migrated provisioner proposal instead of changing old one ------------------------------------------------------------------- Sat Jan 25 14:25:31 UTC 2014 - vuntz@suse.com - Update to version 3+git.1390659907.6aa187e: + Fix test on wrong file + Use die helper + Add helpers to read/write json + Use new helpers to read/write json + Check that /etc/crowbar/provisioner.json contains what we need + Update provisioner proposal with repos in /etc/crowbar/provisioner.json + Fix wrong variable name + Fix ask_yes_no_default_no to work when using logging + Move code finding allocated nodes to a helper function + Add subcommand to reboot nodes + Use die instead of "echo foo; exit 1" + Improve wording + Add a status subcommand, which is now the default ------------------------------------------------------------------- Fri Jan 24 17:46:22 UTC 2014 - vuntz@suse.com - Update to version 3+git.1390585508.d542f97: + Rework user interaction + Do not create data dir from main script + Do not do anything for Hyper-V nodes + Do not call knife to get node attributes + Log why we skip a specific node in pre step ------------------------------------------------------------------- Thu Jan 23 14:35:39 UTC 2014 - vuntz@suse.com - Initial version.