Table 3 lists resolved issues in NorthStar Controller Release 4.2.0. The identifier associated with each entry is the tracking number in the Juniper Networks Problem Report (PR) tracking system.
Table 3: Resolved Issues in NorthStar Controller 4.2.0
In some circumstances, when the connectivity was lost between the servers, a limited number of NorthStar processes may be restarted in all the servers in a NorthStar cluster including on the standby servers. This behavior did not impact the functionality of the active server. However, in order to minimize the confusion, NorthStar HA agent has been updated to avoid this issue.
When dealing with very large scale networks, NorthStar Device Profile took up to 30 seconds to 1 minute to load in the Web GUI. Changes have been made to decrease the loading time to a couple of seconds.
When IPv6 was used to configure the router ID, NorthStar REST API queries reported wrong Segment Routing attributes.
When routers were added in NorthStar Controllers using REST API (routers not coming from the live network via BGP-LS), a duplicate node index was created.
When MTU was not consistently configured in the network, PCE server process could stop responding. Protection has been added to mitigate the impact of network misconfiguration.
When a user updated the MD5 string directly in the Web UI grid cell, versus using the Modify button, PCEP MD5 string for routers was reset in Device Profile. Protection has been added to avoid resetting MD5 key.
When there was MTU mismatch between servers in NorthStar cluster, a communication issue within the Cassandra cluster could trigger a re-election. Additional protection was added to cope with this misconfiguration.
Although it is documented to use NorthStar VIP to launch NorthStar Web UI, it was also possible to launch the web UI on backup servers using the server physical IP. Not using VIP to reach NorthStar Web UI is now blocked.