The release for Project Clearwater sprint “Aragorn” has been cut. The code for this release is tagged as release-118 in GitHub.
This release includes the following bug fixes:
- Sprout converts a 503 Service Unavailable from Homestead into a 404 Not Found (https://github.com/Metaswitch/sprout/issues/1623)
- Calls to +22233 and +5678;rn=+22233 are routed differently when the default route is used (https://github.com/Metaswitch/sprout/issues/1621)
- Sprout crash at high traffic load when configured with a CFS TAS. (https://github.com/Metaswitch/sprout/issues/981)
- Monit does not monitor ralf_uptime or poll_ralf when Ralf has an IPv6 signalling address (https://github.com/Metaswitch/clearwater-monit/issues/52)
- upload_shared_config fails when £ characters are in shared_config (https://github.com/Metaswitch/clearwater-etcd/issues/414)
- etcd memory usage grows unbounded on all nodes, eventually crashing cluster (https://github.com/Metaswitch/clearwater-etcd/issues/399)
Note that in this release, we are changing the way that Geographically Redundant (GR) deployments are set up. Specifically, we were splitting the etcd cluster so that there is one cluster for each site, rather than a single cluster spanning both. Because of this change, there are some one-time extra steps that must be taken prior to upgrading to this release. These steps differ depending on whether you are using a GR deployment or not:
If you are using a non-GR deployment (which is the norm):
Before you upgrade to this release, you must perform the following steps on any one node:
- Add sprout_registration_store=sprout.<zone> to /etc/clearwater/shared_config
- If you are using Ralf nodes, add ralf_session_store=ralf.<zone> to /etc/clearwater/shared_config
- Run cw-upload_shared_config
- Wait for all nodes to pick up the new config by using cw-check_restart_queue_state to check when all nodes have restarted
Once this is done and all nodes have restarted, you can upgrade as normal.
If you are using a GR deployment:
Because of the move to split the etcd clusters across sites, upgrading a GR deployment is not supported using the usual methods. If you want help upgrading a GR deployment, contact us on the mailing list for instructions.
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.