Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

Navigation  Back up to About Overview 
[+] Expand All
[-] Collapse All

No index entries found.

Known Issues

This section lists the known issues in hardware and software in Junos OS Release 15.1F5 for the PTX Series.

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

General Routing

  • Uneven load balancing of traffic may occur if the traffic stream changes only in the bits 0-15 of the Layer3 destination IPv6 address. This limitation may not be visible if the other parameters effecting the load balance change along with L3_DST, such as L3 source IP address, L4 source/destination ports and so on. PR1065515
  • In FPC1and FPC2, ip-l4 and mpls-l4 use different register bits. They are configured enabled/disabled separately in the CLI commands. In FPC3, ip-l4 and mpls-l4 use the same register bit. It can not be configure enabled/disabled separately. PR1075888
  • IPv4/IPv6 Port Mirroring will not be supported on egress interface. PR1081329
  • On FPC3, the strict priority mode scheduling may not be accurate on 10G interfaces with large packet sizes. PR1082439
  • On FPC3, policer counters accuracy is affected because the policer counters is not getting correct data when on chip counters are offloaded to off chip counters. This transaction is not automatic, therefore causing extra packets to be counted. This causes the policer counter to be higher than the actual count.PR1089330
  • The MAC filter limit is reduced to 256 for PTX Series routers with Third Generation FPCs installed. Due to this, the number of IPv6 address on IFLs with subnet mask greater than /104 is reduced to less than 256. Since for such IPv6 addresses with subnet mask greater than /104, the last 24 bits do not remain the same for each IFL. One SNMA is generated for one IPv6 address. PR1098761
  • On FPC3, strict priority mode scheduling may not be accurate with large packet sizes. PR1099387
  • MPLS EXP classifiers for explicit-null labels is not supported on the gladiator platform due to ASIC limitation. This feature will be addressed in the future release. PR1109583
  • "clear services accounting flow" command should not be used in 15.1F4 or 15.1F5 releases on inline Jflow on Gladiator for PTX Series. This command is specific to Jflow and is not supported in these releases. PR1117181
  • If a static route points to a set of interfaces effectively resulting in static route pointing to a unilist nexthop, it is possible that the selector weights may not be initialized correctly resulting in traffic drop. You can mitigate this issue by deactivating and then activating the static route configuration. PR1120370
  • The problem is seen during SNMP polling from two different servers at 45 to 60 PPS while execution of stress commands on DUT NextGen PTX5000 related to AE bundles. PR1130894
  • Under certain conditions in PTX Series platform, the Master Routing Engine may fail to relinquish the mastership upon Control Board (CB) internal switch failure. PR1132557
  • When Fast Re-route is configured with over Aggregate Ethernet bundles, on highly scaled set-ups, upon failure of primary path the switchover to bypass path will happen within a second. Eg: On PTX3000 platform configured with 5000 or more link-protected LSPs, the switchover to bypass path in case of primary path failure, will happen within a second. PR1138402
  • When smart configuration is configured and user does an VMhost regular upgrade, user observes an error print NG_Routing Engine_VMM: "/usr/bin/ Command not found." This can be ignored. This has no impact on upgrade. PR1144881
  • This is a limitation of the avago optics. PR1159999
  • On the P3-15-U-QSFP28 PIC, a low rate of PCS errors may be seen on some ports after rebooting the FPC. Also, for each fifth port (ports 4, 9, and 14) to be active, the corresponding third port (2, 7, and 12) must be active with the same port speed configured. For example, if port 2 is not active, port 4 will not function. PR1163105
  • ISIS may flap during RE switchover. PR1163770
  • On a full box FPC3 chassis, FPC CPU Utilization reaches 100% for a short period of time (between 30 seconds to 1 minute) during AE member links speed changes. PR1163830
  • Traffic may drop during Routing Engine switchover. PR1164107
  • For PTX5000 platform with Broadway based FPCs,BFD V6 sessions with aggressive timers or when CPU load is high on the Routing Engine might flap eg during GRES or if less than 10 mili second timer has been configured. This is applicable only for IPV6 BFD sessions on PTX platform for Broadway Based FPCs. PR1165719
  • When all of the following conditions are met, Forwarding traffic disruption for 2 to 5 minutes may be observed, on a PTX 5k dual Routing Engine system.

    GRES (Graceful Routing Engine Switchover) attribute is configured.

    Two Routing Engines are functional and the router is forwarding traffic as expected.

    No software notification for Routing Engine switchover is issued.

    A Routing Engine, that is running as a master Routing Engine, is unmounted from the chassis, by manually pulling the Routing Engine board from the chassis Routing Engine slot.

    The second Routing Engine takes detects and takes control of the chassis as the new master Routing Engine.

    This is due to a delay in software on the Packet Forwarding Engines in trying to establish a software connection with the new master Routing Engine that is still installed on the chassis. As a result the Packet Forwarding Engines may reboot and would cause forwarding traffic disruption for 2 to 5 minutes. The traffic forwarding however recovers when the Packet Forwarding Engines software boot up and establish a software connection with the new master Routing Engine.PR1165925

  • There is a possibility of Aggregate Ethernet interface going down when master Routing Engine goes down. This issue is intermittently seen.PR1166186
  • In mixed mode when the PVQ configuration is changed while there is traffic flowing, ifds may go down. PR1167076
  • Gladiator: Commit fails after adding member links of same speed from QSFP28 PIC to existing AE bundle created using Universal PIC. PR1167780

High Availability (HA) and Resiliency

  • On MX Series platform with Junos release 15.1R1 or above, while a core dump is in progress, if we try to access the dump directory, due to the deadlock defect, the system might hang and crash. As a workaround, we should not access the "/var/crash" directory till the core dump is complete. PR1087082

Interfaces and Chassis

  • After removing a child link from AE bundle, in the output of "show interface <AE> detail", the packets count on the remaining child link spikes, then if add back the previous child link, the count recover to normal. PR1091425
  • During deactivation of interfaces in a scaled setup the Packet Forwarding Engine may reboot resulting in traffic loss for a short period.
  • FPC3, FIT:ukern core dump occurs at jprds_encap_desc_add/expr_nh_fwd_create_flabel_descr/expr_nh_fwd_create_all _descr with negative events deactivate/activate of mpls configuration ,l2vpn routing instance and FPC reboot. PR1155943
  • ISIS may flap during Routing Engine switchover. PR1163770
  • In very rare chances, a stream drain error message (error code 0x2104ce) might get triggered by frequent interface flap or by port speed changes on the new PTX QSFP PICs. This causes traffic to be dropped. To fix the problem, restart the FPC. PR1169948
  • Traceroute will not work if Gladiator based line card is the ingress line card of a transit router (SWAP router) in 15.1F5.14. PR1171119


  • This is a rare scenario, where LSP's BW is modified to maximum possible value of the link bw in one commit, upon commit some of the lsps will be delayed to signal to new bandwidth. For Ex: While static bw of lsp's is modified by apply-group, let's say 500 lsp's each has has 1m reservation is modified by apply-group configuration to 2m where total link bw is 1g. In this case some of the lsp's will be delayed to signal because of non-availability of bw. These lsps will be signaled when they have sufficient bw to signal. This will only delay the process, does not fail them to resignal with new bw. PR1125323
  • In RSVP-based P2MP scenario, if a sub-LSP switchover to bypass LSP due to PIC offline, then when a new sub-LSP is established via setup-protection, then the deletion of the old sub-LSP might result in deletion of both the sub-LSPs. PR1132050
  • On FPC3, there is a small loss of traffic when there is a NSR switch for static MPLS-LSP. PR1162985

Routing Protocols

  • BFD sessions (both PPMAN and TOE based sessions) using firewall daemon (dfwd) for steering Rx packets to Anchor Packet Forwarding Engine with NSR configured may flap post switchover. This is restricted to PTX Series platform with FPC 3 equipped. The issue is seen only with BFD over AE and Multi-Hop BFD sessions. PR1112218
  • When we have two paths for the same route, the route gets pointed to Unilist NH which in turn gets pointed to two separate Unicast NHs. The route is determined by OSPF and we have BFD enabled on one of the paths which runs through an l2circuit path. When the link on the l2circuit gets cut, the link flap is informed by BFD as well as through OSPF LSAs. Ideally the BFD should inform the link down event before the OSPF LSA. But at the current situation, the OSPF LSAs update the current event a second before BFD. Due to this reason, we do get the route to be pointing to a new UNILIST NH with the weights swapped. But the Unicast NH for which the L3 link is down, gets added to the Unilist NH, the BFD assumes the link to be up, and hence updates the weights inappropriately and hence we do see traffic loss Once the BFD link down event is processed at OSPF protocol level, now the route points to only UNICAST NH and hence we do see traffic flowing through the currently active link The traffic outage would be hardly for less than a second during FRR. Also, this can be avoided if the BFD keepalive intervals are maintained around 50 ms with multiplier as 3 as opposed to 100ms with a multiplier of 3. PR1119253

Related Documentation

Modified: 2016-09-29