3.3.15.72

3.3.15.72

What's new in 3.3.15.72

  • No warning when uploading a license with a serial number mismatch - TB-6199

    • Added a warning during license upload when its serial number does not match the one currently used by the host. The warning is visible in the license installation log on the Web Portal and does not block the installation.

 

  • Media Relay profile options revert to default values when toggling related settings - TB-6661

    • Fixed an issue in the Web Portal where changing the "Allow low-delay media relay" or "RTP security mode" options in a profile's Media Relay settings would reset other fields (such as "Use RTP anchoring") to their default values instead of preserving the saved configuration.

 

  • Loss of service during configuration change with many NAPs under call load - TB-6743

    • Fixed a timing issue during configuration changes that reduce the number of NAPs while calls are in progress. The issue could cause toolpack_engine to restart, interrupting service. Configurations with 300+ NAPs were more likely to be affected.

 

  • Failing calls with WINK_START_ACK CAS signaling - TB-7000

    • Fixed an issue where the ACK wink was not sent after digit collection on the incoming side. Only CAS-to-CAS bridging or loopback configurations were affected.

 

  • Call failure when SIP URI contains long percent-encoded user-part - TB-7014

    • Fixed an issue where SIP URIs with percent-encoded user-parts longer than 63 characters could cause a service disruption. The system now correctly handles URIs up to 127 characters, and URIs exceeding this limit are now rejected with a 481 response.

 

  • Gateway application restarts on TMG under sustained call load - TB-7017

    • Fixed a resource tracking issue in conference call handling. During rapid conference setup and teardown, internal state could become inconsistent, eventually exhausting system resources and forcing the gateway to restart.

 

  • Intermittent SNMP monitoring disruption during adapter reboot - TB-7029

    • Fixed a timing issue in SNMP session management that could cause the monitoring service to restart during an adapter reboot. The service recovered automatically but generated unnecessary diagnostic reports during the interruption.

 

  • REGISTER failures after TCP reconnection for users with long usernames - TB-7050

    • Fixed an issue where the ProSBC could not distinguish registrations from different TCP connections for users with long SIP usernames, causing repeated registration failures.

 

  • Call Trace Export Showing Merged Sessions - TB-7067

    • Fixed an issue where call traces in the Web Portal could display events from two unrelated calls merged into a single trace.

 

  • Support for TLS private key management in Web Portal - TB-7085

    • Added support for uploading TLS private keys alongside certificates through the Web Portal and REST API. Administrators can paste combined PEM content (certificate and private key) into the certificate field or submit it via the REST API. Certificate-only uploads continue to work unchanged. When editing a local TLS certificate, the Web Portal now indicates whether a private key is present.

      Certificate names are now validated to only allow letters, digits, underscores, hyphens, dots, and spaces.

 

  • H.248 IVR tones not always played - TB-7088

    • Fixed an issue where IVR tones (ringback, announcements) could be missing in certain H.248 call flows. The gateway now ensures a reliable audio path for tone playback when an IVR termination is added to a context.

 

  • "Intermediate" certificate type are now "Trusted" certificates - TB-7131

    • The Intermediate certificate type is removed from the Trusted Certificates section in the Web Portal. Existing Intermediate certificates are automatically migrated to Trusted on upgrade. The REST API continues to accept the Intermediate type for backward compatibility.

 

  • Web Portal restarts repeatedly with 500 errors during initial setup - TB-7138

    • Fixed an issue on new deployments where the Web Portal triggered 2 extra restarts per startup cycle, which could disconnect the active session on slower systems