Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

Known Behavior


Learn about known behavior in Junos OS Evolved Release 19.2R1.


  • When firewall filter containing interface-match is applied on output interfaces, match happens for interface on which packets go out (and not for interface on which packets have come in). This behavior would be different from some of the existing platforms. PR1403263

  • Static link protection is not supported on the PTX10003. PR1382692

  • After you insert an optical transceiver, the LINK UP LED flashes more than once during the first 60 seconds. PR1403833

  • If the internal Ethernet switch becomes inaccessible over PCIe, the marvell-cpss-app might encounter an intermittent core when trying to access the bus. Should this occur, the marvell-cpss-app will restart, detect the missing switch during the restart, and fail to configure or access the switch. Traffic in the host path will not be dropped due to the core/restart. PR1405373

  • In its current form, the evo-cda-zx process can consume a relatively high amount (more than RPD) of CPU in idle state. This is related to periodic statistics gathering. PR1412651

  • The FPC offline reason does not show as "Offlined due to major errors" in the show chassis fpc output when the system is brought down due to a CM error infra offline action. PR1412871

  • PTX10003-80C/160C: ~0.4% drops are seen at line rate when packet sizes are below 170b PR1415012

  • When you change the severity for a URI, the old severity error statistics are getting updated under the new severity. PR1417191

  • Restart of certain processes like CDA, AFTMAN, resilD will cause the system to reboot. PR1421993

  • If you apply a filter with a term 1514 applied on full PFEs, the full line rate at the PFE level cannot be achieved from a 168-190 packet type. PR1424351

  • The following XML tags are missing in Junos OS Evolved for the show system buffers CLI: <memory-statistics> <current-mbufs> <cached-mbufs> <total-mbufs> <current-mbuf-clusters> <cached-mbuf-clusters> <total-mbuf-clusters> <max-mbuf-clusters> <packet-count> <packet-free> <current-jumbo-clusters-4k> <cached-jumbo-clusters-4k> <total-jumbo-clusters-4k> <max-jumbo-clusters-4k> <current-jumbo-clusters-9k> <cached-jumbo-clusters-9k> <total-jumbo-clusters-9k> <max-jumbo-clusters-9k> <current-jumbo-clusters-16k> <cached-jumbo-clusters-16k> <total-jumbo-clusters-16k> <max-jumbo-clusters-16k> <current-bytes-in-use> <cached-bytes> <total-bytes> <mbuf-failures> <cluster-failures> <packet-failures> <jumbo-cluster-failures-4k> <jumbo-cluster-failures-9k> <jumbo-cluster-failures-16k> <sfbuf-requests-denied> <sfbuf-requests-delayed> <io-initiated> PR1429626

  • There are differences in xml tags for cli "show system directory-usage." PR1429629

  • For cli show system uptime, there are output xml tag differences compared to Junos. PR1429636

  • The following XML tags are missing in Junos OS Evolved for the show system virtual-memory CLI: <system-virtual-memory-information> <vmstat-memstat-malloc> <vmstat-memstat-info> <memstat-name> <inuse> <memuse> <high-use> <memstat-req> <memstat-size> <vmstat-memstat-zone> <vmstat-zone-info> <zone-name> <zone-size> <count-limit> <used> <free> <zone-req> <vmstat-sumstat> <cpu-context-switch> <dev-intr> <soft-intr> <traps> <sys-calls> <kernel-thrds> <fork-calls> <vfork-calls> <rfork-calls> <swap-pageins> <swap-pagedin> <swap-pageouts> <swap-pagedout> <vnode-pageins> <vnode-pagedin> <vnode-pageouts> <vnode-pagedout> <page-daemon-wakeup> <page-daemon-examined-pages> <pages-reactivated> <copy-on-write-faults> <copy-on-write-optimized-faults> <zero-fill-pages-zeroed> <zero-fill-pages-prezeroed> <transit-blocking-page-faults> <total-vm-faults> <page-faults-requiring-io> <pages-affected-by-kernel-thrd-creat> <pages-affected-by-fork> <pages-affected-by-vfork> <pages-affected-by-rfork> <pages-cached> <pages-freed> <pages-freed-by-deamon> <pages-freed-by-exiting-proc> <pages-active> <pages-inactive> <pages-in-vm-cache> <pages-wired-down> <pages-free> <bytes-per-page> <total-name-lookups> <positive-cache-hits> <negative-cache-hits> <pass2> <cache-deletions> <cache-falsehits> <toolong> <vmstat-intr> <vmstat-intr-info> <intr-name> <intr-cnt> <intr-rate> <vm-kernel-state> <vm-kmem-map-free> PR1429639

  • The following XML tags are missing in Junos OS Evolved for the show system queues CLI: <queues-statistics> <interface-queues-statistics> <interface-queue> <name> <octets-in-queue> <max-octets-allowed> <packets-in-queue> <max-packets-allowed> <number-of-queue-drops> <protocol-queues-statistics> <protocol-queue> PR1433449

  • More than one link-flap can be seen within 60 seconds, when DAC optics is jacked out and jacked-in, before the link stabilizes. PR1438770

  • NETCONF GET for JTI telemetry is not supported on PTX10003-80C PR1439175

General Routing

  • RIP and RIPng is currently not supported on PTX10003-80C PR1437544

  • The show/clear igmp snooping hierarchy option missing in Junos OS Evolved. PR1439755


  • Traceroute fails to list the complete paths b/w spoke CEs w/ and w/o no-vrf-propagate-ttl configured. PR1436638

Network Management and Monitoring

  • For commit scripts in Junos OS Evolved, tracing is always-enabled (like flight recorder mode). Therefore, the command hierarchy for enabling tracing options for commit scripts in Junos OS Evolved has been removed. There will be no updated command to replace the old one. PR1423291

  • If you use IPv6 via two different user-specified routing instances to configure two remote servers belonging to different VRFs, Junos OS Evolved forwarding syslog messages fail. The messages only fail when you change the configuration several times during setup. The syslog messages do not fail for manual cases. PR1436314

Routing Policy and Firewall Filters

  • When an firewall filter with counters action is binded on AE interface, and AE interface doesnt have any child member links, counters of filter are not shown in "show firewall" cli command. PR1433501

User Interface and Configuration

  • If management instance is configured and then you do a load override of another configuration (which has management instance in a different VRF), the management instance gets deleted. When committing this, the commit hangs indefinitely. This happens as the management instance now moves to a different VRF and the current connectivity to the device is lost. This is the expected behavior. The user has to log back in again and then the device will work normally. PR1421947

  • If you commit "set interfaces ae0 gigether-options 802.3ad ae0" without checking "Invalid configuration," the command will commit without an error message. This may cause an ifmand process core. PR1426018