JN0-352 Questions and Answers

The intent of this filter is to match and discard traffic from one single, specific host MAC address, which requires the source-mac-address match condition to reference the complete, exact 48-bit address using an explicit /48 mask (or no mask at all, which Junos treats as an implicit full match) so that only that one precise address is matched. As configured, the term instead specifies a /40 mask on 00:12:34:56:00:07, meaning only the first 40 bits (the first five octets, 00:12:34:56:00) are actually significant to the match, while the sixth octet is effectively wildcarded across its full range; this produces an imprecise, overly broad match criterion rather than the single-host match the administrator intended, and is a common configuration mistake that undermines the predictability and correctness of MAC-based filtering. The lesson tested here is that source and destination MAC address match conditions in Junos firewall filters must be specified with the correct, deliberate, and explicit prefix length appropriate to the intended scope — typically /48 for a single exact host — rather than an arbitrary or miscalculated mask value. Firewall filters under family ethernet-switching are fully capable of matching on both source-mac-address and destination-mac-address, which rules out that option entirely. Layer 2 MAC-based filters function correctly on both access and trunk interface modes, so interface type is not a limiting factor here. There is no indication in the exhibit that the filter's application direction (input versus output) is reversed. Reference topics: Junos Enterprise Switching – Layer 2 Security, MAC Address Match Conditions and Prefix Length Precision.

The destination address 172.16.200.5 falls within the 172.16.0.0/16 aggregate's summarized range but does not fall within either of the two active, more-specific contributing routes actually present in the table — 172.16.10.0/24 and 172.16.99.0/24 — meaning no route more specific than the /16 itself exists to cover it. Under Junos's longest-match forwarding logic, the lookup for 172.16.200.5 therefore resolves to the aggregate route itself, and since aggregate routes are installed by default with a reject next hop rather than any real forwarding path, the router drops the packet and simultaneously generates an ICMP destination-unreachable message back to the originating source, explicitly signaling that no valid path exists for that specific destination even though a covering summary is being advertised. This reject behavior is precisely why aggregate routes are valuable for reducing the number of routes advertised upstream while still providing clear, immediate feedback for traffic aimed at address space within the summary that has no genuine underlying route, rather than silently black-holing it or misdirecting it toward an unrelated contributing route's next hop. The aggregate route itself remains active and installed throughout this process — a gap in contributing coverage does not deactivate the aggregate, since the aggregate's activation depends only on at least one contributing route being active, which is satisfied here by both existing /24 blocks. There is also no default route present in this scenario to fall back upon. Reference topics: Junos Enterprise Routing – Protocol Independent Routing, Aggregate Route Reject Behavior for Uncovered Address Space.

The Complete Sequence Number PDU (CSNP) is the IS-IS mechanism used to summarize the entire contents of a router's link-state database, listing the LSP ID, sequence number, checksum, and remaining lifetime for every LSP the sending router currently believes should exist in the database for that level. On broadcast (LAN) circuits, the elected Designated Intermediate System periodically transmits CSNPs to all routers on the segment as a database-synchronization mechanism; any receiving router compares the summarized list against its own local database, and any discrepancy — an LSP it is missing entirely, or one for which it holds an older sequence number — triggers a targeted request for the specific, updated LSP content. On point-to-point circuits, a single CSNP is exchanged once when the adjacency first comes up, serving the identical synchronization purpose without needing periodic repetition. The Partial Sequence Number PDU (PSNP), by contrast, is used to acknowledge receipt of specific LSPs or to explicitly request specific LSPs identified as missing or outdated after a CSNP comparison, making it a partial, targeted PDU rather than a complete database listing. The LSP itself is the PDU that actually carries the detailed topology and reachability information being advertised, rather than a database summary. The IS-IS Hello PDU (IIH) is used purely for neighbor discovery and adjacency maintenance and carries no link-state database summary information whatsoever. Reference topics: Junos Enterprise Routing – IS-IS, CSNP and Link-State Database Synchronization.

Nonstop active routing's defining architectural feature is that the backup Routing Engine runs a fully live, active instance of the routing protocol daemon (rpd) in parallel with the primary Routing Engine's rpd instance, continuously, from the moment NSR and its prerequisite GRES are established — not merely at the instant a switchover begins. As protocol adjacencies form, routes are learned, and the RIB is updated on the primary Routing Engine, that same protocol and RIB state is mirrored in real time to the backup Routing Engine's rpd process through an internal synchronization mechanism, so the backup's protocol state remains a live, up-to-date replica of the primary's at essentially every moment during normal operation. When a switchover occurs, the backup Routing Engine simply assumes the primary role with protocol adjacencies, routing tables, and forwarding state already fully populated and current, allowing it to continue operating without re-establishing neighbor relationships or re-learning the network topology, which is what makes the switchover genuinely nonstop and invisible to neighboring routers. This is fundamentally different from graceful restart, where the restarting router must rebuild its state by requesting help from cooperating neighbors after the fact. The Packet Forwarding Engine plays no role in synchronizing protocol-layer information such as LSAs or BGP updates; that is strictly an rpd-to-rpd function coordinated through the Routing Engine synchronization infrastructure. Reference topics: Junos Enterprise Routing – High Availability, Nonstop Active Routing Architecture and rpd Synchronization.

This scenario describes end hosts — virtualization hosts running software switches or virtual bridging stacks — unexpectedly generating and transmitting BPDUs on ports that are supposed to be pure edge (host-facing) ports where no BPDUs should ever legitimately be received. BPDU protection (BPDU guard) is the Junos feature purpose-built for exactly this situation: when applied to an edge-designated interface, it continuously monitors that port for any incoming BPDU, and the moment one is detected, it immediately and automatically disables the port, preventing the unexpected BPDU source from ever participating in or influencing the spanning-tree topology calculation, thereby protecting the stable, intended topology from disruption caused by unauthorized or accidental BPDU generation at the access edge. This directly and precisely resolves the described problem at its source, since the root cause is illegitimate BPDUs arriving on ports that were never meant to see them. Increasing the max-age timer only adjusts how long stale topology information is retained before being aged out and has no effect on preventing rogue BPDUs from being processed in the first place. Disabling RSTP outright on access interfaces would eliminate spanning tree's loop-prevention protection on those ports entirely, introducing a much more severe risk of undetected Layer 2 loops. Root protection (root guard) defends specifically against a port attempting to become the root port by receiving a superior BPDU claiming a better path to root; it addresses a different threat model than blocking any BPDU outright on a true edge port. Reference topics: Junos Enterprise Switching – Spanning Tree Protocols, BPDU Protection on Edge Interfaces.

Layer 2 MAC learning on an EX Series switch, as on any standard Ethernet bridge, is a passive, frame-driven process rather than an active polling or query-based mechanism. Every time a frame arrives on any interface, the switch inspects the frame's source MAC address field and, if that address is not already present in the Ethernet switching (bridge) table for the associated VLAN, creates a new dynamic entry recording the source MAC address, the VLAN it was learned on, and the specific ingress interface through which the frame arrived. Junos also stamps each learned entry with an aging timestamp, which is refreshed every time a subsequent frame from that same source MAC is received; entries that receive no refreshing traffic within the configured MAC aging interval (300 seconds by default) are eventually purged from the table to keep it current and to reclaim table space from hosts that have moved or gone offline. This source-address-driven, per-frame learning process happens continuously and automatically with no dependency on spanning-tree state changes — although a topology change notification can trigger accelerated aging of the table to flush potentially stale entries faster, learning itself is not gated by or contingent upon such an event occurring. The switch never learns addresses from the destination field of a frame, since the destination address, if already known, is used purely for the forwarding lookup decision, not for populating new table entries. Reference topics: Junos Enterprise Switching – Layer 2 Switching Fundamentals, Source MAC Learning and Table Aging.

Once the two candidate routes have equal route preference and equal local preference (both defaulting to 100 unless modified by policy), the Junos BGP algorithm proceeds to its third decision step, which compares the length of each candidate's AS path attribute and selects whichever path carries the shortest AS path, since a shorter path is interpreted as traversing fewer autonomous system hops to reach the origin. This step occurs well before the algorithm ever reaches the specific EBGP-over-IBGP preference rule, which is only evaluated later, after local preference, AS path length, origin type, and MED have all already been compared and found equal between the candidates; because an IBGP-learned copy of a route typically retains the identical AS path that was originally received from the external AS (IBGP does not prepend the local AS number the way crossing an AS boundary does), it is entirely possible, and often the deciding scenario, for AS path length to differ between competing paths well before the algorithm would ever need to fall back on the IBGP-versus-EBGP tiebreak. Selecting a route purely because it was 'learned through IBGP' is not a rule that exists anywhere in the algorithm and would in fact typically be the opposite of Junos's later EBGP-preference tiebreak. Neither next-hop IP address value nor router ID value is compared with a 'highest wins' rule at any point in the standard path selection sequence — where router ID is used at all, Junos prefers the lowest value, not the highest, making both of those distractors doubly incorrect. Reference topics: Junos Enterprise Routing – BGP, Path Selection Order: AS Path Length Precedes EBGP/IBGP Preference.

The Junos BGP path selection algorithm proceeds through a strict, ordered sequence of comparison steps, and once next-hop reachability and route preference (administrative distance) have been confirmed as equal, local preference is the very first BGP-specific path attribute evaluated in the decision process. A route carrying a higher local preference value is always chosen over a competing route to the same prefix with a lower local preference, and this comparison takes absolute precedence over every subsequent attribute in the algorithm; if local preference values differ between candidate paths, the selection process concludes at this step without ever needing to examine AS path length, origin type, or MED at all. Only when two or more competing paths share an identical local preference value does the algorithm proceed to its third step, comparing AS path length and preferring the shortest, and only after a further tie does it move on to origin type (preferring IGP-origin over EGP-origin, and EGP-origin over incomplete), followed by MED comparison (preferring the lowest value, and only among routes learned from the same neighboring AS by default). Because local preference is evaluated strictly before all three of the other listed attributes, it is the attribute BGP consults first among the choices given, making it the deciding factor in scenarios such as this exhibit where multiple autonomous systems advertise competing paths toward the same 192.168.1.0/24 prefix. Reference topics: Junos Enterprise Routing – BGP, Understanding BGP Path Selection Order.

A link aggregation group bundles multiple physical Ethernet links into a single logical interface (an ae interface) specifically to provide both increased aggregate bandwidth and link-level redundancy. By default, the aeX interface remains operationally up as long as at least one member link within the bundle remains active and passing traffic; the LAG does not require every member link to be functional to continue operating. When one of the four 1-GbE member links in this scenario fails, LACP (if in use) or the underlying link-status monitoring detects the failure, removes that specific link from the active forwarding set, and redistributes traffic hashing across the three remaining healthy member links automatically and without operator intervention. End devices and upper-layer protocols continue to see a single, continuously up ae interface throughout this process, experiencing only a proportional reduction in available aggregate bandwidth rather than any interruption in connectivity. This default behavior is precisely why LAG is a foundational building block for resilient EX Series switch interconnects in enterprise and data center designs, since it allows individual optic, cable, or transceiver failures to be absorbed transparently. None of the distractor behaviors — full administrative shutdown of the bundle, forwarding limited to broadcast traffic only, or a temporary disabling of all member links during renegotiation — reflect actual Junos LAG behavior, which is designed explicitly to avoid any such disruption. Reference topics: Junos Enterprise Switching – Layer 2 Switching, Link Aggregation Group Resiliency.

On EX Series switches, aggregated Ethernet (ae) interfaces do not exist by default; before any ae interface can be created, referenced, or brought into service, an administrator must first reserve the desired number of aggregated Ethernet device slots by configuring set chassis aggregated-devices ethernet device-count number at the [edit chassis] hierarchy. This statement allocates the internal resources needed for that many ae interfaces to exist as valid, configurable objects; until it is committed, any interface-level configuration referencing ae0 (such as the aggregated-ether-options and family ethernet-switching stanzas shown in the exhibit) is accepted syntactically by the CLI but never actually instantiates a live ae0 device, which is exactly why show interfaces ae0.0 returns 'device ae0 not found' despite the interface appearing fully configured. Increasing the device-count to at least 1 (or higher, to cover the number of LAG bundles planned) immediately resolves this and allows ae0 to appear as an operational interface once its member links are also associated with it. This is a foundational and frequently overlooked prerequisite step in LAG deployment on Juniper switches. There is no separate licensing requirement for aggregated Ethernet functionality on EX Series switches, a reboot is not needed since the chassis statement takes effect upon commit, and the LACP mode (active versus passive) is unrelated to whether the ae interface object itself exists in the first place. Reference topics: Junos Enterprise Switching – Link Aggregation, Configuring the Aggregated Ethernet Device Count.

Voice VLAN is the Junos feature specifically designed to allow a single access port to simultaneously handle two distinct traffic streams: untagged data traffic from a PC or workstation using the port's normal access VLAN, and separately tagged voice traffic generated by an IP telephone that is typically daisy-chained between the switch port and the connected PC. When voice VLAN is configured on an access interface, the switch instructs the attached IP phone (commonly via LLDP-MED) which VLAN ID to tag its voice traffic with, and the switch is then able to correctly classify and prioritize that tagged voice traffic into its designated voice VLAN while simultaneously continuing to treat any untagged frames arriving on that same physical port as belonging to the port's regular, untagged access VLAN — precisely satisfying the requirement to accept both tagged and untagged traffic together on one access port. Native VLAN, by contrast, is a trunk-port-only concept that designates which VLAN untagged frames on a trunk should be associated with; it has no applicability to an access port and does not enable simultaneous tagged traffic acceptance there. Default VLAN simply refers to the VLAN an interface belongs to absent other configuration and carries no dual-traffic capability. IRB interfaces provide Layer 3 gateway functionality for a VLAN and are unrelated to how tagged versus untagged frames are accepted at Layer 2 on a given physical port. Reference topics: Junos Enterprise Switching – VLANs, Configuring Voice VLAN on Access Interfaces.

The described symptom — excessive packet drops specifically attributed to exceeding the MTU on a GRE tunnel carrying full-size video streaming traffic across default-MTU gigabit Ethernet interfaces — is a direct consequence of the 24 bytes of GRE encapsulation overhead pushing packets that were already at or near the standard 1500-byte Ethernet MTU past the physical interface's transmission limit once the outer GRE and IP headers are added. The most durable fix is to proactively increase the physical member interfaces' MTU beyond the 1500-byte default — to 1524 bytes or higher — which provides sufficient headroom to absorb the full GRE encapsulation overhead without ever causing the fully encapsulated packet to exceed the interface's transmission capability, allowing full-size 1500-byte payloads to pass through the tunnel intact and without fragmentation or drops. As a complementary measure, configuring the clear-dont-fragment-bit option on the tunnel or member interfaces instructs Junos to clear the DF (don't fragment) bit on oversized packets rather than dropping them outright when they do exceed the available MTU, allowing the Packet Forwarding Engine to fragment and later reassemble the traffic instead of silently discarding it, which directly addresses the reported packet loss even under transient conditions. Simply raising the MTU only to 1500 bytes fails to account for the 24-byte GRE overhead and would not resolve the underlying oversize condition. Routing the tunnel over a higher-bandwidth link addresses potential congestion but does nothing to correct an MTU-based packet-size violation, which is the root cause explicitly identified in this scenario. Reference topics: Junos Enterprise Routing – Tunneling, GRE MTU Sizing and Fragmentation Handling for High-Throughput Applications.

OSPF neighbor state progression, as defined in RFC 2328 and displayed by Junos through show ospf neighbor, moves through a well-defined set of named states: Down, Attempt, Init, 2-Way, ExStart, Exchange, Loading, and Full. Down is the initial state, indicating no Hello packets have been recently received from a potential neighbor on that interface; it is a genuine, valid OSPF state and is correctly included here. Init indicates that a Hello packet has been received from the neighbor, but that neighbor has not yet listed the local router's own router ID within its Hello packet, so bidirectional communication has not yet been confirmed; this too is a standard, valid state. Loading occurs after the Exchange state, during which Database Description packets have been fully exchanged and the router is now actively requesting the specific, more-detailed LSAs it identified as missing or outdated from its neighbor via Link State Request packets; it is likewise a standard, valid OSPF neighbor state. 'Established' is not an OSPF state at all — that terminology belongs to BGP's finite state machine, where Established represents the fully operational session state; the OSPF equivalent concept is instead called Full. 'Reject' is not a recognized OSPF neighbor state under any circumstance and does not appear anywhere in the RFC 2328 state machine, making it an invalid distractor as well. Reference topics: Junos Enterprise Routing – OSPF, The OSPF Neighbor State Machine.

The factory-default storm control configuration on EX Series switches operates in a purely rate-limiting mode rather than an interface-shutdown mode. When broadcast, multicast, or unknown-unicast traffic on a port exceeds the configured bandwidth-percentage threshold, the default action is simply to police the offending traffic type and silently drop only the packets that exceed that threshold, while all traffic within the allowed rate continues to be forwarded normally and the interface itself remains administratively and operationally up throughout the event. This explains the observed symptom precisely: during the burst of unknown-unicast frames, the switch dropped the excess volume above the configured limit, producing a brief period of packet loss, but as soon as the burst subsided and traffic fell back under the threshold, storm control simply stopped dropping packets and normal forwarding resumed automatically — no administrative shutdown, recovery timer, or manual intervention was ever involved, because the interface was never taken down in the first place. This default drop-only behavior is distinct from the optional interface-shutdown action, which must be explicitly configured if an administrator wants storm control to disable a port outright on violation (as seen in a related scenario requiring the clear ethernet-switching recovery-timeout command). Storm control by default covers broadcast, multicast, and unknown-unicast traffic types together, not multicast alone, and it has no functional relationship to GRES, which governs Routing Engine redundancy rather than Layer 2 traffic policing. Reference topics: Junos Enterprise Switching – Storm Control, Default Drop Behavior Versus Interface Shutdown Action.