Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 
 

Known Limitations

Learn about limitations in this release for the PTX10001-36MR, PTX10003, PTX10004, and PTX10008.

For the most complete and latest information about known Junos OS Evolved defects, use the Juniper Networks online Junos Problem Report Search application.

EVPN

  • If a packet with an unknown inner ether-type is received at the device over an EVPN-MPLS tunnel, then such a packet is dropped. PR1564431

General Routing

  • The license details are not fetched in broadcast mode when the device and the license server are in the same subnet. As a workaround, configure license server explicitly on the the device using the command set system license server address ip so as to not use broadcast mode. PR1463584

  • Line rate traffic can be achieved on a PTX10008 router at 256B with 400K routes scale. PR1468280

  • MAC entries are stuck in l2ald or l2alm for some time after the clear ethernet switching command is executed. PR1507812

  • The excess-rate configuration in port schedulers might not be completely honored in certain scenarios, where there could be cases where the output queues configured with guaranteed-rate with no excess-rate might end up getting more share in the excess region of the bandwidth than the queues with explicit excess-rate configured. PR1528124

  • When a scheduler-map binding is removed from an interface, then the default scheduler-map is bound to the interface. If default scheduler-map is an oversubscribed scheduler map for the interface, then that map is not applied to this interface and all the "interface queue" counters for this interface show statistics as 0. PR1539052

  • PTX10008: By default IPv6 addressing is configured with /64 subnet irrespective of the subnet configured on the DHCP server side. PR1539839

  • Indirect next-hop acknowledgment configuration statement makes the rpd request forwarding next-hop acknowledgment from all the Packet Forwarding Engines before doing indirect next-hop change to point to the new forwarding next hop. However, this acknowledgment is not 100% guaranteed. The acknowledgment might not work if the same forwarding next-hop (which the indirect next hop is pointing to) installation is attempted through a route add or change path because in the route add or change flow, the next-hop acknowledgment is not enabled. Since the next-hop installation is already attempted through route flow, indirect next-hop change flow cannot request acknowledgment on a forwarding next hop which is already in queue to be sent to the the Packet Forwarding Engine. One possible workaround might be to enable acknowledgment for all the next hops. But this may have scaling implications and performance penalty. PR1547432

  • Egress sflow sampling with sampling ratios lower than 1:1 (e.g. 1:10, 1:100, etc.) could result in up to 17% reduction in sampling capacity. This issue is not applicable to ingress sflow sampling. Also, if sampling ratio of 1:1 is configured for egress sampling, the issue is not seen. PR1567647

  • In case of fixed form factor chassis like PTX10003 and PTX10001-36MR, the samples on all the interfaces on all the virtual FPCs (if any) are considered to make the adaptive sampling decision. Due to this, the maximum adaptive sampling rate supported is 35000 samples/sec per device and not per-FPC in the case of fixed form factor chassis platforms. PR1577229
  • After UDP tunnel header decapsulation, MPLS Payload with MPLS label 0 and Label 2 is dropped because this requires the decap of MPLS label and lookup of inner payload which is not supported without loopback in BT ASIC based products. We support only scenarios where we decapsulate tunnel header and forward the packet based on the exposed MPLS Label. PR1580641

  • PTX10003-80C or PTX10003-160C interface queue and voq does not report drops when the low priority queue is slightly oversubscribed ( seeing page timeouts). PR1581490

Interfaces and Chassis

  • When there is over-subscription traffic across Packet Forwarding Engines in PTX10003-80C or PTX10003-160C, and one of the Fabric ASICs (ZF) goes bad, the software takes an automatic action to recover the system by issuing an automatic SIB offline or online. However, if the egress traffic is at close to line-rate, the traffic would take time to converge till which there would be fabric drops. PR1580376

MPLS

  • The rpd process might crash after network service configuration changed (like changing the range of MPLS labels) without rebooting all the Routing Engines (which is a system mandatory step). PR1461468

Routing Protocols

  • BGP does not advertise route with Entropy Label Capability and load-balance-label is not as stitching point on a specific platform. PR1529387