Changelogs

Update Cassandra 4.1 gc.log configuration options to reduce logging printed information and to allow analysis by censum tool. (#1161334)

Updated rvconfig set-desired-running-state command so it lowercases instance names for MDM instance IDs (as SIMPL/MDM do) (#994044)

Initconf sets directory and file permissions to the primary user (instead of root) when extracting custom data from yaml configuration files. (#510353)

Add new charging option 'cap-ro' to support mixed CAMEL and Diameter Ro deployment. (#701809)

Add support for configuring multiple destination realms for Diameter Ro. (#701814)

Updated example configuration for conference-mrf-uri to force TCP (#737570)

Corrected the SNMP alarm that was previously monitoring totalFree memory, it now checks for availReal memory instead. (#853447)

Modified the validation scripts to avoid checking rhino liveness & alerts when IPSMGW is disabled. (#737963)

Allow upload config if there is no live node for a given VM type (#511300)

Cassandra 4 container upgraded to 4.1.3 (#987347)

Updated system package versions of libwebp, bind, bpftool, kernel, open-vm-tools, perf, and python to address security vulnerabilities. (#1023775)

Upgraded Apache Tomcat to 8.5.95 version. (#986693)

The minimum supported version of SIMPL is now 6.13.3. (#290889)

TSN upgrades are supported when all other non-TSN nodes are already upgraded to 4.1.3-1.0.0 or higher. Refer to Major upgrade from 4.0.0 of TSN nodes (#290889).

TSN VM supports 2 Cassandra releases - 3.11.13 and 4.1.1; the default is 4.1.1 for new deployments, 3.11.13 can be selected by setting the custom-options parameter to cassandra_version_3_11 during a VM deployment. New rvtconfig cassandra-upgrade allows one-way switch from 3.11.13 to 4.1.1 without outage. Refer to Cassandra version switch procedure for TSN nodes (#290935)

New rvtconfig backup-cds and rvtconfig restore-cds commands allow backup and restore of CDS data. Refer to Take a CDS backup (#290889)

New rvtconfig set-desired-running-state command to set the desired state of non-TSN initconf processes. Refer to Resume Initconf in non-TSN nodes(#290889)

Fixed a race condition during quiesce that could result in a VM being turned off before it had completed writing data to CDS. (#733646)

Improved the output when rvtconfig gather-diags is given hostname or site ID parameters that do not exist in the SDF, or when the SDF does not specify any VNFCs. (#515668)

Fixed an issue where rvtconfig would display an exception stack trace if given an invalid secrets ID. (#515672)

rvtconfig gather-diags now reports the correct location of the downloaded diagnostics. (#515671)

The version arguments to rvtconfig are now optional, defaulting to the version from the SDF if it matches that of rvtconfig. (#380063)

There is now reduced verbosity in the output of the upload-config command and logs are now written to a log file. (#334928)

Fixed service alarms so they will correctly clear after a reboot. (#672674)

Fixed rvtconfig gather-diags to be able to take ssh-keys that are outside the rvtcofig container. (#734624)

Fixed the rvtconfig validate command to only try to validate the optional files if they are all present. (#735591)

The CDS event check now compares the target versions of the most recent and new events before the new event is deemed to be already in the CDS. (#724431)

Extend OutputTreeDiagNode data that the non-TSN initconf reports to MDM based on the DesiredRunningState set from rvtconfig. (#290889)

Updated system package versions of nss, openssl, sudo, krb5, zlib, kpartx, bind, bpftool, kernel and perf to address security vulnerabilities. (#748702)

Hazelcast on SMO and SGC nodes now only binds to the clustering interface, not all interfaces. (#716776)

Enable rvtconfig to upload the config containing differences or errors when there is no node alive without using skip_diff. (#738214)

Added REM Local instance to MAG nodes for backward compatiblity with REM provisioning requests. (#829927)

Updated SMO configuration so that the server for UE-Reachability notifications listens on the internal interface only. (#667685)

Fixed an issue where rvtconfig validate would fail on a SMO node when the IP-SM-GW was disabled. (#737963)

Added support for parallel upgrades (except for TSN, SMO and SGC). Refer to Notes on parallel vs sequential upgrade for more details.

Added support for VM Recovery. Depending on different situations, this allows you to recover from malfunctioning VM nodes without affecting other nodes in the same VM group.

Added a low-privilege user, named viewer. This user has read-only access to diagnostics on the VMs and no superuser capabilities. (OPT-4831)

Migrated XCAP server from REM in Tomcat to the new XCAP service in Rhino on the MAG node. (VOLTE-9819)

Updated VM sizing for TSN and MMT nodes. The old sizes are retained for upgrades, but new sizes should be used for new deployments. (#370050)

Access to VMs is now restricted to SSH keys only (no password authentication permitted). (OPT-4341)

The minimum supported version of SIMPL is now 6.10.1. (OPT-4677, OPT-4740, OPT-4722, OPT-4726, #207131) This includes different handling of secrets, see Secrets in the SDF for more details.

Made the system-notification-enabled, rhino-notification-enabled, and sgc-notification-enabled configuration options mandatory. Ensure these are specified in snmp-config.yaml. (#270272)

MMT and MAG nodes no longer use a clustering interface.

Changed prefix of internal XCAP hostname in DNS template from internal-xcap. to xcap.internal.. XCAP requests are now required to have a Host starting with xcap. to conform to the format outlined in the 3GPP TS 23.003 specification. (OPT-4714)

Settings in the product-options section of the SDF now raise an error if specified in high-level YAML (as they once used to be), rather than simply warning the user that they have moved. (OPT-3380)

Made mag-vmpool-config.yaml xcap-domains value format more strict. Each value in xcap-domains must now start with xcap. (including the .). Previously they only needed to start with xcap (without the .). This stricter check brings it more in-line with the XCAP (and BSF) domain name format outlined in the 3GPP TS 23.003 specification. (OPT-4352)

Removed NAF filter configuration related to local nonce storage — only cassandra-based nonce storage is supported now. Also removed NAF filter-specific Cassandra connection details as Cassandra connection is now via the Cassandra CQL RA, the same as the BSF. (VOLTE-9616)

The Local instance from REM is no longer functional, and will be removed in the next maintenance release. To manage the MAG nodes in REM, use the host-specific instance instead (normally RVT-mag.DC1-<hostname>).

Improved ShCM security by automatically restricting ShCM API access only to nodes that need it - MMT, SMO & MAG. (VOLTE-10193)

Added support for Rhino Node IDs to be configurable for unclustered VMs, and made this field mandatory. (#235644)

The SGC JMX management API now only listens on localhost. (OPT-4834)

SMO/SGC VMs now refuse to apply OCSS7 configuration changes unless the documented SGC reconfiguration procedure is followed. (#276422)

Added a list of expected open ports to the documentation. (OPT-3724)

Added enter-maintenance-window and leave-maintenance-window commands to rvtconfig to control scheduled tasks. (OPT-4805)

Added a command liveness-check to all VMs for a quick health overview. (OPT-4785)

Added a command rvtconfig report-group-status for a quick health overview of an entire group. (OPT-4790)

Split rvtconfig delete-node-type into rvtconfig delete-node-type-version and rvtconfig delete-node-type-all-versions commands to support different use cases. (OPT-4685)

Added rvtconfig delete-node-type-retain-version command to search for and delete configuration and state related to versions other than a specified VM version. (OPT-4685)

Added rvtconfig calculate-maintenance-window to calculate the suggested duration for an upgrade maintenance window. (#240973)

Added rvtconfig gather-diags to retrieve all diags from a deployment. This has been optimised to gather diags in parallel safely based on the node types alongside disk usage safety checks. (#399682, #454095, #454094)

Added support for Cassandra username/password authentication. (OPT-4846)

system-config.yaml and routing-config.yaml are now fully optional, rather than requiring the user to provide an empty file if they didn’t want to provide any configuration. (OPT-3614)

Added tool mdm_certificate_updater.py to allow the update of MDM certificates on a VM. (OPT-4599)

Added support for configuring SAS connectivity using DNS hostnames. (OPT-4716)

Added support for configuring REM debug logging through high level config. (OPT-4799)

The VMs' infrastructure software now runs on Python 3.9. (OPT-4013, OPT-4210)

All RPMs and Python dependencies updated to the newest available versions.

Updated the linkerd version to 1.7.5. (#360288)

Updated the Cassandra version to 3.11.13. (#255349)

Upgraded PostgreSQL to version 12. (OPT-5022)

Use the Microsoft build of OpenJDK instead of the CentOS OpenJDK. (#255345)

Fixed issue with default gateway configuration.

initconf is now significantly faster. (OPT-3144, OPT-3969)

Added some additional clarifying text to the disk usage alarms. (OPT-4046)

Ensured tasks which only perform configuration actions on the leader do not complete too early. (OPT-3657)

Tightened the set of open ports used for SNMP, linkerd and the Prometheus stats reporter. (OPT-4061, OPT-4058)

Disabled NTP server function on the VMs (i.e. other devices cannot use the VM as a time source). (OPT-4061)

The report-initconf command now returns a meaningful exit code. (DEV-474)

Alarms sent from initconf will have the source value of RVT monitor. (OPT-4521)

Removed unnecessary logging about not needing to clear an alarm that hadn’t been previously raised. (OPT-4752)

Authorized site-wide SSH authorized public keys specified in the SDF on all VMs within the site. (OPT-4729)

Reduced coupling to specific SIMPL VM version, to improve forwards compatibility with SIMPL. (OPT-4699)

Moved initconf.log, mdm-quiesce-notifier.log and bootstrap.log to /var/log/tas, with symlinks from old file paths to new file paths for backwards compatibility. (OPT-4904)

Added the rvt-gather_diags script to all node types.

Increased bootstrap timeout from 5 to 15 minutes to allow time (10 minutes) to establish connectivity to NTP servers. (OPT-4917)

Increase logging from tasks which run continuously, such as Postgres and SSH key management. (OPT-2773)

Avoid a tight loop when the CDS server is unavailable, which caused a high volume of logging. (OPT-4925)

SNMPv3 authentication key and privacy key are now stored encrypted in CDS. (OPT-3822)

Added a 3-minute timeout to the quiesce task runner to prevent quiescing from hanging indefinitely if one of the tasks hangs (OPT-5053)

The report-initconf command now reports quiesce failure separately to quiesce timeout. (#235188)

Added a list of SSH authorized keys for the low-privilege user to the product options section of the SDF. (#259004)

Store the public SSH host keys for VMs in a group in CDS instead of using ssh-keyscan to discover them. (#262397)

Add mechanism to CDS state to support forward-compatible extensions. (#230677)

Logs stored in CDS during quiesce will be removed after 28 days. (#314937)

The VMs are now named "Metaswitch Virtual Appliance". (OPT-3686)

Updated mmt-gsm-stats.xml stats config to use SCCCamelToIMSReoriginationIN SBB parameter set. (OPT-4137)

Fixed issue where SAS config was required on SMO nodes even if the IPSMGW function was disabled. (OPT-3614)

Enabled SAS on the MAG node — NAF/BSF/XCAP requests will now be reported to the configured SAS. (OPT-4255)

Removed the rvtconfig split-sdf and shcm-properties-to-yaml tools that are no longer required for upgrades. (OPT-3395)

Fixed issue where replication would be enabled for call types where it is not supported. (VOLTE-10172)

Update MAG nginx config to add X-Ua-OpenSSL-Cipher-Suite header to xcap server requests containing UE-nginx SSL connection cipher. (#340633)

Corrected a defect that could cause the uplevel SGC/SMO VM to be unable to rejoin the cluster following an upgrade. (#230582)

Fixed bug where Rhino SLEE state would not be preserved across a Rhino restart. (RHI-6378)

Ignored exceptions thrown when attempting to start the SLEE immediately after it has been started. (OPT-4019)

During upgrades, make the Rhino quiesce stage more efficient by splitting up tasks, and allowing the leader node more time to perform Postgres connection management. Allow 3 attempts to upload logging before continuing. (OPT-4859)

Fixed rare cases where initconf could become stuck on a call to the flushconfiguration command. (OPT-4810)

Ensured Rhino only listens for management commands on the management interface. (OPT-4934)

Moved all Rhino logs to /var/log/tas/rhino. (OPT-4906)

Changed permissions on the log files rhino.log and alarms.csv, so that all users - notably the viewer user - can now read them. (OPT-4906)

Fixed rare failure to converge when a node was booting while a different node was configuring an RA. (OPT-4927)

Updated postgres stop systemd timeout timer to from 60 minutes to 5 minutes. (OPT-5050)

Rhino will no longer restart or stop when PostgreSQL is restarted or stopped via systemctl. (#260638)

Failure to stop Rhino during quiesce will no longer prevent quiesce completing. (#296266)

Removed saving/restoring of OID mappings on nodes running Rhino. OID mappings are static now in Rhino. (OPT-4144)

Fixed an issue around enabling SNMP v3 in Rhino before the credentials had been configured. (#233470)

Removed the seeds_allocation Cassandra table. (OPT-4786)

Added support for repair and cleanup scheduling. Repairs are scheduled for peers after a Cassandra node has been assassinated. Cleanups are scheduled for peers after a TSN recovery. All scheduled actions are run sequentially on each peer node to avoid parallel invocation.

All TSN nodes now check for keyspace availability and schema version agreement. (OPT-5060)

Store Cassandra logs in CDS after quiesce. (OPT-4693)

The CDS replication factor is now proportional to the number of TSN nodes in the cluster. If there are more than 5 TSN nodes in the cluster the replication factor is capped at 5. (OPT-4823)

Increase TSN commissioning timeout to 30 minutes from 20 minutes. (#308746)

Alarms are now raised when the ramdisk Cassandra partition reaches 80% and 90% usage (instead of 90% and 95%). (#338936)

The replication factor is now correctly set to 1 on Cassandra keyspaces in lab deployments with one or two TSN nodes. (#187408)