Support Support Downloads Knowledge Base Case Manager My Juniper Community

Knowledge Base

Search our Knowledge Base sites to find answers to your questions.

Ask All Knowledge Base Sites All Knowledge Base Sites JunosE Defect (KA)Knowledge BaseSecurity AdvisoriesTechnical BulletinsTechnotes Sign in to display secure content and recently viewed articles

20.4R2-S2-EVO: Software Release Notification for JUNOS Software Version 20.4R2-S2-EVO

0

0

Article ID: TSB18101 TECHNICAL_BULLETINS Last Updated: 06 Jul 2021Version: 1.0
Alert Type:
SRN - Software Release Notification
Product Affected:
PTX and QFX series running Junos Evolved
Alert Description:
Junos Software Service Release version 20.4R2-S2-EVO is now available for download from the Junos software download site
Download Junos Software Service Release:
  1. Go to Junos Platforms - Download Software page
  2. Input your product in the "Find a Product" search box
  3. From the Type/OS drop-down menu, select Junos SR
  4. From the Version drop-down menu, select your version
  5. Click the Software tab
  6. Select the Install Package as need and follow the prompts
Solution:

Junos Software service Release version 20.4R2-S2-EVO is now available.

20.4R2-S2-EVO - List of Fixed issues
PR Number Synopsis Category: "agentd" software daemon
1590432 Non zero values might be displayed against the drop field in ?show network-agent statistics? CLI post switchover scenarios.
Product-Group=evo
In case of switchover scenarios, if the collector which was connected to older master, tries to connect to new master immediately, non-zero values could be seen in drops field for ?show network-agent statistics? CLI. These are not actual packet drops. Each packet sent as part of streaming data would contain a header which would have a meta information of the packet contents. One such field in the header indicates the current packets sequence number. This is a monotonically increasing number for each packet from a producer of telemetry data. During switchover cases, collectors may receive initial packets with a higher sequence number which could get reset to 0 after sometime. Due to this pattern, the cli would show non zero values against drops field. Note: These are not actual packet drops and there is no functionality impact. However it is not expected to see further increase in this value shown against the drops field.
PR Number Synopsis Category: Border Gateway Protocol
1591717 Doing BGP disable/enable in a short time interval on a scaled NSR router can result in backup RPD restart
Product-Group=evo
In an NSR enabled router with high BGP scale (peer scale as well as RIB scale), when BGP is disabled and enabled in a short interval of time such that the BGP enable happens before BGP completes the cleanup for the earlier disable, then master RPD and backup RPD can get out-of-sync for the 'BGP enable' synchronized configuration commit. The backup RPD restarts itself to recover from this out-of-sync condition.
PR Number Synopsis Category: Tracking PD specific issues on BRCM platforms running EVO
1577375 QFX5130 and QFX5220 Object anomalies seen with PTP TC config
Product-Group=evo
On QFX5130 and QFX5220, if PTP is configured and then deleted, an object anomaly will be seen, of an object that is not deleted. With each cycle of PTP delete/create, an additional object will remain, thus leaking memory of that object.
PR Number Synopsis Category: Express BT PFE L3 Features
1538918 EXP rewrite might not take effect on the exposed label during performing PHP
Product-Group=evo
On the BT chip based platforms with MPLS EXP rewrite rules configured, If the router is performing PHP (Penultimate Hop Popping), the EXP rewrite will not take effect on the exposed label. Due to this, packets may get into the wrong queue and even get dropped during a network congestion situation.
1589803 Traffic loss observed on global repair after disable of active path forcing FRR
Product-Group=evo
In a scaled configuration with most of the next-hop on the device, egressing out of one [physical/aggregate] port and unitlists are created with ports of two FPCs, transient traffic loss may be observed when active path is disabled. When active patch is disabled, churn happens in the above-mentioned next-hops and Transit MPLS routes. Until all this reprogramming completes, PTX series routers drop traffic.
1593079 After RE switch-over, error messages associated : JexprSlowCntrRead - Unable to get the plct Inst for pfeIdx: 255, User-type: OVFM_OFFCHIP_NEXTHOP_CNTR
Product-Group=evo
This error message is only seen after RE switchovers. After RE SWO, few of counters may not have accurate stats.
1593548 Load-balance may not take effect for the L2VPN traffic on the PTX10008 platforms
Product-Group=evo
On the PTX10008 platforms running 20.4R2 onwards, load-balance may not take effect for the L2VPN traffic on the FPCs after configuring/deleting the knob "no-vlan-tagging".
PR Number Synopsis Category: EVO Netstack DDoS (ddosd and JTD)
1588556 RE policing status is updated correctly in the ddos-protection show command outputs
Product-Group=evo
RE policing is not applicable for these platforms. However, show command outputs such as "show ddos-protection statistics" and "show ddos-protection protocols parameters" reflected RE policing to be enabled all the time. RE policing knob is updated correctly, so that the correct output is displayed now.
PR Number Synopsis Category: EVO Netstack FIB Service Daemon
1594041 The LDP sessions might not come up on EVO based platforms
Product-Group=evo
The LDP session might remain in down state after the FPC restart or switchover is being performed, on EVO based platforms. The frequency of occurrence of the issue might depend upon how often the fibd restarts and it is more likely to be seen on dual-RE systems.
PR Number Synopsis Category: Issues related to EVO interface statistics.
1575623 Huge Invalid stats shown in show interface statistics when an interface is being removed and added from the AE bundle
Product-Group=evo
Huge Invalid stats shown in show interface statistics when an interface is being removed and added from the AE bundle and clear interface statistics CLI command is issued in quick succession.
PR Number Synopsis Category: Index management related issues; idmd* daemons
1578635 PTX10004 :: PDT - BOWMORE - with NSR enabled, after fpc0 restart, traffic fail over to fpc2 but loss due to dlu.ucode.not_routable, irp.core.trapcode.cfg_err, sw.irp_nh_discard_sample counters increased
Product-Group=evo
In a scale configuration with most of the nexthops in DUT, egressing out of one [physical/aggregate] port and unilists are created with ports of two FPCs, transient traffic loss is observed when remote FPC is rebooted. In this test case, the scale numbers are ~23K unicast nexthops, ~24K unilist nexthops and ~18K Aggregate nexthops, ~24K Transit MPLS routes, ~1M IPv4 routes and ~215K IPv6 routes. When remote FPC is rebooted , churn happens in the above mentioned nexthops and Transit MPLS routes.
PR Number Synopsis Category: software upgrade infra issues
1595470 Some TCP sessions might not be established after performing the "request system snapshot" command
Product-Group=evo
On all Junos OS Evolved (EVO) platforms, if performing the "request system snapshot" command, the device will add 128.0.0.0/2 route with eth interface in the main table, which will cause the SYN-ACK packets to go to the wrong interface. So the device may not reply SYN-ACK when receiving TCP SYN messages and TCP-based protocols will be affected.
PR Number Synopsis Category: EVO Netstack Juniper Tunnel Driver Module
1587748 TCP server with low TCP MSS causes TCP SESSION not to be established on Junos EVO platforms
Product-Group=evo
On all Junos EVO platforms, when any TCP(Transmission Control Protocol) server which originates a TCP session is Junos EVO device, the TCP MSS(maximum segment size, the MSS refer to the Linux kernel MSS) with larger size than what it should have selected, the TCP packets will be dropped by the network. At which point there will be no further packets delivered to the TCP peer. Take BGP protocol for example, the BGP session will fail to receive a BGP keep-alive and will flap the connection. If the protocol change to SSH, the SSH session may just appear to "hang".
PR Number Synopsis Category: Sonic-Agent bugs, issues that are part of EVO
1600932 config interface ip remove is not working correctly
Product-Group=evo
using config interface ip remove command to remove ip configuration on an interface is not working correctly. The workaround is to use config_db.json file and do config reload -y. Procedure is: config interface ip remove .... config interface ip add .... config save config reload -y
PR Number Synopsis Category: Configd, ffp issues
1534858 The firewall filter for both IPv4 and IPv6 might not work when it is applied through apply-groups.
Product-Group=evo
On Junos OS Evolved PTX devices, the firewall filter for both IPv4 and IPv6 does not work when it is applied through apply-groups on the subinterface
PR Number Synopsis Category: Express PFE MPLS Features
1586122 Packet loss may be seen during global repair of FRR
Product-Group=evo
In a high scaled setup with the combination of Unicast Nexthops, Unilist nexthops, composite nexthops and Aggregate nexthops and when majority of the unicast nexthops egress out of one physical port, packet loss is observed during Global repair of FRR events.
PR Number Synopsis Category: Port Mirroring feature on express PFE
1529413 Global Port-Mirroring -Applyed with Deactivation Doest not display XML Correctly for ptx10003
Product-Group=evo
When global port-mirroring configured on DUT and queried for xml info, everything works as expected. When that global port-mirroring configuration is deactivated and xml is queried, display info is missing port-mirroring-instance info. Expected output: ============== [edit] Actual Output: ============= [edit] What is Missing? ==============
1593276 20.4R2-BT: Port mirroing instance down with mirrored output as tagged interface(Old PR1524869).
Product-Group=evo
Local Port-Mirroring doesn't work on interfaces with non-zero unit.
PR Number Synopsis Category: PFE infra to support jvision for EVO
1592468 Duplicate jvision leaf "oper-status" tag for IFD index 16386 have mismatch value
Product-Group=evo
With the jvision feature enabled on PTX10008, for oper-status for IFD index 16386 (which is an internal interface and is not configurable), there are 2 issues: 1. We see duplicate values exported for this leaf 2. These values have a mismatch between the RE and PFE exported values
PR Number Synopsis Category: Platform infra to support jvision
1595103 On EVO platforms, "type" leaf value for "FPC3:PIC0:PORT0:XcvrX" displays XCVR as opposed to TRANSCEIVER displayed in JUNOS
Product-Group=evo
On EVO platforms, "type" leaf value for "FPC3:PIC0:PORT0:XcvrX" displays XCVR as opposed to TRANSCEIVER displayed in JUNOS
PR Number Synopsis Category: Protocol Independant Multicast
1591685 PIM joins might not be synchronized between master and backup REs because of ppmd restart
Product-Group=evo
In an NSR enabled router, PIM (S, G) joins might go out of synchronization when "request system application app ppmd node re0 restart" is performed.
PR Number Synopsis Category: PTX10K specific platform PRs
1597076 Power budget test case : FPCs not coming up if PSMs inserted after switchover
Product-Group=evo
Post switchover if "show chassis hardware" shows entries for removed PSMs then reinserting these PSMs will not power back the offlined FPCs.
PR Number Synopsis Category: Bugs related to SONiC on PTX10K platform
1590983 The orchagent and syncd processes might crash while deleting routes
Product-Group=evo
On PTX platforms with SONiC release in a scaled ECMP use case, SONiC cores can occur while deleting routes and the ECMP group.
PR Number Synopsis Category: EVO Services Sflow PRs for defect & enhancement requests
1592788 On PTX10008 and PTX10001-36MR platforms, sflow sample-rate configuration greater than 16000000 not supported
Product-Group=evo
On PTX10008 and PTX10001-36MR platforms, sflow sample-rate configuration greater than 16000000 not supported
 

20.4R2-S2-EVO - List of Known issues
PR Number Synopsis Category: EVO RCB software
1592258 For ptx10001-36mr the "Host 0 Voltage Threshold Crossed" alarm will be present.
Product-Group=evo
For ptx10001-36mr, the software driver reads the voltage threshold erroneously causing the "Host 0 Voltage Threshold Crossed" alarm to present on the device.
PR Number Synopsis Category: PFE L2 forwarding features on BT based platforms
1601915 EVO:JDI_FT_REGRESSION:PLATFORM:VSCAPA:L2CP : After deleting old vlan and adding new vlan,lag interface is not learn by static vlan
Product-Group=evo
In ptx10008, at times, MVRP enabled trunk ports can go into blocked/designated state if the peer connected to the interface has no vlans configured in its trunk port.
PR Number Synopsis Category: Express BT PFE L3 Features
1590656 Minor Traffic drop will be seen during MBB of RSVP LSP without optimize-adaptive-teardown knob
Product-Group=evo
Minor TransientTraffic drop will be seen during MBB of RSVP LSP without optimize-adaptive-teardown knob
1593086 [PDT] AftCore: AftValidate: Validation failed for token:18446744073709551615,error during interface flaps HALT test case.
Product-Group=evo
When add/delete AE child links, this error message might appear. it means that when inserting AftExprNode (AE for Multicast), Afttoken for some AE child links is not ready. It failed in AftCore validation. AftExprNode won't be created at this stage. It should not have any functional impact. When installing Multicast Composite Nexthop, AftExprMcAgg token will be installed if not exist.
PR Number Synopsis Category: software upgrade infra issues
1580374 FPC is stuck in onlining state and seen continuously rebooting during ISSU
Product-Group=evo
During a restart based upgrade (request system software add restart), if an FPC is offline at the start of the upgrade, the upgrade flow will transition the FPC to online. With this issue, the routing engine will fail to add this FPC to the allowed upgrade members, and (1) FPC will continuously reboot itself waiting for the upgrade to be done, and (2) Upgrade flow will not converge waiting for all nodes to be ready.
PR Number Synopsis Category: Interface PRs defect & enhancement requests
1595179 [EVO] Licenses used value for Port Bandwidth Usage (PAYG) is not updated when new interfaces are created, unless we restart IFMAND application.
Product-Group=evo
The licenses used value for "Port Bandwidth Usage (PAYG)" field under "show system license" does not update when new interfaces are created. In order to correct the value, the IFMAND process needs to be restarted to invoke a refresh.
PR Number Synopsis Category: EVO L2 Control Protocols Support
1592473 L2cpd-agent may go unresponsive on starting telemetry service if LLDP/STP is configured
Product-Group=evo
L2cpd-agent may go unresponsive on starting telemetry service if LLDP/STP is configured.
PR Number Synopsis Category: EVO linux defects & enhancement requests
1564506 Uncompressed vmcore with no backtrace
Product-Group=evo
vmcore is corrupted without valid backtrace
PR Number Synopsis Category: Express PFE MPLS Features
1592997 "show mpls lsp ingress statistics" commands times out once the lsp are up after clear rsvp session all trigger
Product-Group=evo
This might happen when the system is under churn after running "clear mpls lsp" and is downloading route/NH. Since aftmand-bt is busy downloading routes and nexthops, servicing CLI commands can get deprioritized.
PR Number Synopsis Category: PFE infra to support jvision for EVO
1592505 jvision: Duplicate leaf under /interfaces/interface/state tag tree
Product-Group=evo
With the jvision feature enabled on PTX10008, the following leaves display duplicate values: - carrier-transitions tag - high-speed tag These are under available the path :/interfaces/interface/state heirarchy.
PR Number Synopsis Category: Label Distribution Protocol
1594405 rpd core in backup RE@ in mirror_process_recvd_data_queue with mldp NSR config
Product-Group=evo
RPD core occurs most of the times in standby LDP module when static LDP p2mp config is deleted.
1596395 MLDP transit traffic statistics on EVO are resetting after RE switchover
Product-Group=evo
MLDP transit statistics reset after RE switchover. Dev is debugging and a fix will be provided in release post 21.2R1
PR Number Synopsis Category: KRT Queue issues within RPD
1586466 Error message "RPD_KRT_KERNEL_BAD_ROUTE" seen on certain scenarios when rpd restart/GRES when NSR enabled which has no functional impact.
Product-Group=evo
During RPD restart and GRES scenario, when an interface is deleted and the corresponding nexthop is about to be deleted , we can see the following error message" RPD_KRT_KERNEL_BAD_ROUTE: krt unsolic client.: lost ifl 0 for route" is thrown . This message does not have any functional impact. This is a corner case and the probability of occurrence is low. The error happens when rpd is re-playing the route whose nexthop's interface was just delinked and linked to a local interface. This message is a transitionary in nature and part of the common code that handles all scenarios. Identifying this specific transition period and avoiding this message needs to be taken up in the next release.
PR Number Synopsis Category: RPD route tables, resolver, routing instances, static routes
1593445 Increase remnant-holdtime according to fib scale and configurations.
Product-Group=evo
The default remnant hold time is 300seconds which is not sufficient for this scale(1.7M fib routes, 37K LSPs, jflow enabled). We need to increase this hold time to 600seconds to avoid any protocol flap during RPD crash or restart. "set routing-options forwarding-table remnant-holdtime 600"
PR Number Synopsis Category: Resource Reservation Protocol
1592305 Transit HIPRI < Protected> LSP traffic GR Loss in the range of 50-200+ msec on Link Failures with LSP self-ping
Product-Group=evo
In a scale configuration when one of the AE interface is disabled, observed 28ms of e2e traffic-loss with default LSP ingress timers. End-to-End Traffic-loss depends scale and capacity of the routers participating in MBB. In this test case, the scale numbers are ~54K Unilist NHs, ~131K composite nexthops, ~33K IPv4 routes, ~86K MPLS routes and ~21K IPv6 routes. LSP ingress timers can be tuned to get zero loss.
PR Number Synopsis Category: PTX10K platform specific fabric PRs
1539685 PFEs on FPC stuck in "READY" state on New Master RE with GRES switch while FPC is restarting.
Product-Group=evo
Software doesn't support GRES, when FPCs are triggered for restart. GRES with FPC triggered for offline or online is supported.
1562066 Traffic drops in certain high scale SRV6 traffic streams when FPCs are offlined with traffic
Product-Group=evo
There is traffic loss on some SRV6 streams post FPC restart. > Is the problem seen all the time with every restart? Yes > Is this seen only with FPC restart? Or are there any other triggers This is the only known trigger for this issue. > How to identify and resolve the problem when it happens Traffic drops will be seen, along with the following error reported in "show system errors active detail" * intr_req_src_ppfe_drop_err e.g., Error Name : dp_1_zfi_1_intr_req_src_ppfe_drop_err Identifier : /fpc/1/evo-cda-bt/0/cm/0/btchip/2/dp_1_zfi_1_intr_req_src_ppfe_drop_err Description : dp_1_zfi_1_intr_req_src_ppfe_drop_err State : enabled Scope : pfe Category : functional Level : minor Threshold : 10 Error limit : 30 Occur count : 2 Clear count : 1 Last occurred(ms ago) : 24095915 Workaround: FPC restart or router restart.
1584706 SIB state may not get updated properly when its physically removed from chassis during swo
Product-Group=evo
SIB state may not updated properly when its physically removed from chassis during RE switchover.
PR Number Synopsis Category: PTX10K specific platform PRs
1478951 PTX10008 graceful OIR leads to missing of sensor data for power and temperature. Follow the optimized step given as a workaround to have OIR working.
Product-Group=evo
PTX10008 graceful OIR leads to missing sensor data for Power and Temperature. Please follow the optimized step given as a workaround to have OIR working.
1591661 FRUs go to Failed state on backup RE after unexpected GRES
Product-Group=evo
Symptom: Unexpected switchover to other RE without a CLI command or application failure causes FRUs to move to "Failed" state. Workaround: On unexpected switchover, reboot the new backup RE using ?request node reboot re[0|1]? Problem Description: On unexpected mastership switchovers (not initiated by CLI or application failure), kernel connector in the new backup RE is not notified of mastership change and therefore FRUs can move to failed state as access fails. As workaround, Please reboot new backup RE after unexpected switchovers.
PR Number Synopsis Category: PTX10K RE EVO Issues
1585845 Scapa RCB: EFT: Unable to power-on RCB0 post node power-off, while following the specific given sequence
Product-Group=evo
To power-off RE and power-on of RE node: To power-off RE node: $ request node power-off re Ensure the RE node is completely powered down. CB state is transitioned to offline. Use the below CLI commands to verify the states before triggering the power-on command. {MASTER} root@re1-re1> show system nodes Node: re0 Node Id : 0 Node Nonce : 0 Status : offline, configured-powered-off. <-- Final state of the RE node. Attributes : {MASTER} root@lazurite-plat-ft-c-re1> show chassis environment cb CB 0 status: State Offline Standby <-- Final state of the CB. Bus Revision 247 FPGA Revision 267 To power-on RE node: $ request node power-on re

 
Modification History:
First publication 2021-07-06
Comment on this article > Affected Products Browse the Knowledge Base for more articles related to these product categories. Select a category to begin.

Getting Up and Running with Junos

Getting Up and Running with Junos Security Alerts and Vulnerabilities Product Alerts and Software Release Notices Problem Report (PR) Search Tool EOL Notices and Bulletins JTAC User Guide Customer Care User Guide Pathfinder SRX High Availability Configurator SRX VPN Configurator Training Courses and Videos End User Licence Agreement Global Search