The release for sprint "For Whom The Bell Tolls" has been cut. The code for this release is tagged as release-73 in Github.
This release contains the culmination of our work over the last few sprints to make configuration and clustering of Clearwater deployments simpler. Now, new nodes can be configured with just their local IP configuration and the IP address of one or more existing nodes and they will automatically configure themselves to match, and join their clusters. For more information, see http://clearwater.readthedocs.org/en/latest/Automatic_Clustering_Config_Sharing We’ve also upgraded to use Cassandra 2.0.14 rather than 1.2.12. This release also includes the following fixes. · An I-CSCF without S-CSCF doesn't support ENUM (https://github.com/Metaswitch/sprout/issues/1060) · Sprout doesn't respond to OPTIONS polls with public_hostname in the request URI (https://github.com/Metaswitch/sprout/issues/1043) · scscf_uri should be created using bracket_ipv6_address.py (https://github.com/Metaswitch/sprout/issues/1028) · Bono doesn't use RFC 3263 DNS resolution to locate Sprout (https://github.com/Metaswitch/sprout/issues/1021) · Sprout occasionally crashes during termination (https://github.com/Metaswitch/sprout/issues/963) · Can't add the /etc/clearwater/remote_cluster_settings file without restarting sprout (https://github.com/Metaswitch/sprout/issues/871) · We don't always generate an explanatory SAS log for failures (https://github.com/Metaswitch/sprout/issues/800) · Requests that fail because the local cassandra isn't running aren't retried (https://github.com/Metaswitch/homestead/issues/221) · Chronos can't bind to localhost (https://github.com/Metaswitch/chronos/issues/128) · Chronos doesn't use its default configuration (https://github.com/Metaswitch/chronos/issues/126) · nginx site isn't re-enabled after upgrade (https://github.com/Metaswitch/crest/issues/210) · Homer/Homestead-Prov installs require Mock, but this hasn't been provided as an egg (https://github.com/Metaswitch/crest/issues/209) · Potential race condition when using clearwater-upgrade to update Homer (https://github.com/Metaswitch/crest/issues/177) · NameError exception in DnsRecordsManager when scaling up (https://github.com/Metaswitch/chef/issues/180) · chef-client needs to be nicer (https://github.com/Metaswitch/chef/issues/75) · poll_cassandra.sh should default to localhost if cassandra_hostname not set (https://github.com/Metaswitch/clearwater-cassandra/issues/30) · The diags monitor doesn't collect cassandra or memcached information for memento (https://github.com/Metaswitch/clearwater-infrastructure/issues/171) · NTP daemon is not monitored (https://github.com/Metaswitch/clearwater-infrastructure/issues/154) · Makefile does not spot changes in cpp-common (https://github.com/Metaswitch/clearwater-snmp-handlers/issues/41) · Statistics don't work if some statistics don't change at least once every 5 seconds (https://github.com/Metaswitch/clearwater-snmp-handlers/issues/28) · Makefile uses wildcard dependencies (https://github.com/Metaswitch/clearwater-snmp-handlers/issues/21) · Manual install instructions are sadly lacking (https://github.com/Metaswitch/memento/issues/33) · clearwater-nginx fails to install because nginx already started (https://github.com/Metaswitch/clearwater-nginx/issues/17) · `monit unmonitor memcached` also unmonitors `poll_memcached`... (https://github.com/Metaswitch/clearwater-monit/issues/16) · Astaire's init.d script doesn't set namespace_prefix so it doesn't work when using multiple networks (https://github.com/Metaswitch/astaire/issues/19) · HEAT templates should police input parameters (https://github.com/Metaswitch/clearwater-heat/issues/3) · The astaire log cleanup script isn't run because it's not executable (https://github.com/Metaswitch/astaire/issues/21) · Chronos doesn't restart on upgrade (https://github.com/Metaswitch/chronos/issues/132) Upgrading to this release is slightly more complicated than usual. Firstly, there is some release preparation to do. · On any memento nodes (either standalone node, or nodes where Memento is collocated with Sprout), temporarily disable the disk space cron job by running 'chmod 000 /etc/cron.hourly/memento_disk_usage' · Also on any memento nodes, copy the /etc/clearwater/cluster_settings file to /etc/clearwater/memento_cluster_settings. · On any nodes that run Cassandra (Homer, Homestead and Memento), take a backup of the cassandra configuration files in /etc/cassandra. · Update the cassandra.yaml on any nodes running Cassandra to remove any options that are invalid under 2.0.14. If you fail to remove an option, Cassandra won't start after upgrade. It will write a log to /var/log/cassandra/system.log indicating what setting should be removed - you should check that log, remove the relevant settings from cassandra.yaml, and Cassandra should then start up normally. You can then upgrade your deployment. As usual, please follow the instructions at http://clearwater.readthedocs.org/en/latest/Upgrading_a_Clearwater_deployment/index.html. · On any node that runs Cassandra, run ‘nodetool upgradesstables’ to update the on-disk Cassandra files. · The Cassandra 2.0.14 install will install Java 7. If you want to use this as the default (matching the configuration on new installs), optionally run `update-alternatives --set java /usr/lib/jvm/java-7-openjdk-amd64/jre/bin/java`. · If the node is a memento node, re-enable the disk space cron job by running 'chmod 755 /etc/cron.hourly/memento_disk_usage' Finally, you can migrate to the simpler configuration model by following the instructions at http://clearwater.readthedocs.org/en/latest/Migrating_To_etcd. If you are deploying an all-in-one node, the standard image (http://vm-images.cw-ngv.com/cw-aio.ova) has been updated for this release. Ellie
_______________________________________________ Clearwater mailing list [email protected] http://lists.projectclearwater.org/listinfo/clearwater
