Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 
 

Known Limitations

This section describes issues and limitations present in Juniper Cloud-Native Router release 23.2.

Known Issues and Limitations in Juniper Cloud-Native Router Release 23.2

  • JCNR-4348: IPv4 and IPv6 traceroute via JCNR is not working with bgp unnumbered underlay.—When JCNR is deployed between two provider edge routers, the IPv4 and IPv6 traceroute does not work with BGP underlay. In case of BGP unnumbered, interfaces don’t have v4 or v6 routable addresses. So, JCNR should generate TTL expiry error packets with source address as vrf loopback address. However, currently JCNR only supports one loopback in global vrf.

  • JCNR-4346: The vrouter deployment fails in L3 mode if Networkmanager is enabled for the interfaces assigned to JCNR—The vRouter deployment fails in L3 mode with the following error message: error creating TAP interface bond0: Invalid argument. The Networkmanager must be disabled for the interfaces assigned to JCNR and the bond interface must be manually created. Please review the System Requirements for more detail.

  • JCNR-4342: The vRouter does not persist latest interface configuration on pod restart or crash—If interfaces have IP configuration prior to JCNR deployment and post deployment, the interfaces are configured via the cRPD, the vRouter does not persist the latest configuration on pod restart or crash. The latest interface configuration must be manually applied via the cRPD after the vRouter restarts.

  • JCNR-4171: Ping between IPv6 kernel pod interfaces intermittently fails—Ping between IPv6 kernel pod interfaces with JCNR deployed as a CNI in L3 mode fails intermittently with Destination unreachable: Address unreachable error message. It is recommended to disable ipv6 forwarding globally on the host as a workaround.

  • JCNR-4145: Link Local Addresses do not persist for non-default VRFs upon cRPD restart—The link local addresses do not persist for non-default VRFs upon the cRPD restart, if the Networkmanager is enabled for interfaces assigned to JCNR. The Networkmanager must be disabled for interfaces assigned to JCNR at the time of deployment. Please review the System Requirements for more detail.

  • JCNR-4248: Keepalived cores are generated when VRRP is initially configured in cRPD— When VRRP is initially configured in cRPD or deleted and re-configured, keepalived cores are generated in the /var/crash directory.

  • JCNR-4364: PersistentVolumeClaims (PVCs) and PersistentVolumes (PVs) created by JCNR are not deleted upon uninstallation on AWS EKS deployments—When Helm installs the JCNR charts, including the cRPD statefulset, any PVCs and PVs created are not deleted upon JCNR uninstallation. This is standard Helm behavior. The following additional steps are recommended to clean up JCNR PVCs on EKS when uninstalling:

    Note that the JCNR PVCs are in the jcnr namespace, while the PVs have StorageClass set to jcnrstorage.

  • JCNR-4070: Core path not updated as per values.yaml— The core path is not updated as defined under the corePattern attribute in values.yaml. The core_pattern value must be defined on the host before deploying JCNR. Please see System Requirements for more detail.