F5CAB3 Questions and Answers

SSH is a TCP Layer 4 protocol. Applying an HTTP profile causes BIG-IP to expect HTTP headers, breaking SSH sessions. Removing the HTTP profile allows raw TCP forwarding.

SSH is a Layer 4 TCP-based protocol that operates on TCP port 22 and does not use HTTP in any capacity. In the exhibit, the Virtual Server is configured with an HTTP Profile applied, which is inappropriate for SSH traffic and causes connection failures.

According to the BIG-IP Administration: Data Plane Configuration documentation:

An HTTP profile must only be applied to Virtual Servers handling HTTP or HTTPS traffic.

When an HTTP profile is attached, BIG-IP expects HTTP headers and attempts to parse application-layer data.

Non-HTTP protocols such as SSH, FTP (control), SMTP, and other raw TCP services will fail if an HTTP profile is enabled.

Why the other options are incorrect:

A. Set Protocol to UDPSSH uses TCP, not UDP. Changing the protocol would break SSH entirely.

B. Set Source Address to 10.1.1.2The source address setting controls client access restrictions and is unrelated to protocol parsing issues.

C. Set Destination Address/Mask to 0.0.0.0/0The destination address is already valid for a specific SSH service and does not impact protocol handling.

Correct Resolution:

The BIG-IP Administrator should remove the HTTP Profile (set it to None) so the Virtual Server functions as a pure Layer 4 TCP service, allowing SSH connections to pass through successfully.

X-Forwarded-For inserts the original client IP into HTTP headers, preserving client identity while SNAT is enabled.

The configuration provided utilizes Priority Group Activation in conjunction with the min_active_members setting. Priority groups allow an administrator to define primary servers and " backup " servers within the same pool. The BIG-IP prioritizes traffic based on the assigned priority number, with the highest number receiving traffic first.

In this specific configuration, the priority 3 group is the primary group. Therefore, connections are first distributed to all pool members with priority 3 as long as they are available. The system will continue to use only the priority 3 group unless the number of available members in that group falls below the min_active_members value, which is set to 2.

If the priority 3 group has fewer than two active members, the BIG-IP " activates " the next available priority group (priority 2) and distributes traffic among the remaining members of priority 3 and all members of priority 2. This cascading logic continues down the list. Consequently, if both the priority 3 group and the priority 2 group have fewer than two members available, traffic is directed to the priority 1 group. This ensures that even in a multi-server failure scenario, the system has a last-resort group of servers to handle the traffic.

Option D is incorrect because if only one member of priority 3 goes down, there are still two members active (10.12.10.8 and 10.12.10.9). Since 2 is not less than the min_active_members threshold of 2, the priority 2 group will not yet be activated. Option B is incorrect because traffic flows from high priority to low priority, not the other way around.

Comprehensive and Detailed Explanation From BIG-IP Administration Data Plane Configuration/Course Guide/topics:

The monitor is explicitly sending an HTTP/1.1 request. For HTTP/1.1, the request must contain a valid Host header identifying the intended virtual host. The configured monitor sends the URI and Connection: Close header but omits Host: www.f5.com. A web server using name-based virtual hosting can therefore reject the request with HTTP 400 Bad Request , which explains why every pool member is marked down.

F5 ' s official BIG-IP monitor documentation gives the HTTP/1.1 convention in the form " GET / HTTP/1.1\r\nHost: server.com\r\nConnection: close\r\n\r\n " .

Option A follows that required structure: the request line identifies HTTP/1.1, Host: www.f5.com selects the correct hosted site, Connection: Close requests connection termination after the response, and the final \r\n\r\n properly terminates the HTTP header section. F5 also specifies that when HTTP/1.0 or HTTP/1.1 is used, the monitor evaluates the HTTP status code; a non-200 response such as 400 causes the monitored resource to be considered unavailable.

Reference Topics: BIG-IP LTM Monitors Reference → HTTP Monitor Send String , HTTP/1.1 request formatting , pool-member health monitoring .

The BIG-IP Configuration Utility (GUI) organizes information hierarchically to allow for granular management of application objects. A Pool is a collection of backend servers (pool members) that provide the same service. To verify the specific network parameters—such as the IP address and the service port—of the servers within a pool, the administrator must navigate to the specific pool ' s configuration.

The standard procedural path to verify this is Local Traffic > Pools: Pool List, where the administrator selects the specific pool name. Once inside the pool ' s configuration, the Members tab displays a list of all IP addresses and service ports associated with that pool. Under the " Service Port " column, the administrator can confirm if the member is listening on port 80 (HTTP).

Options A and B are incorrect for this specific verification task. While Nodes (Option A) show the health of a physical server, a node represents only an IP address and does not have a " Service Port " associated with it until it is defined as a pool member. Verifying the Health Monitor (Option B) would tell the administrator how the system is checking the member ' s status, but it does not definitively show the port on which the member is actually receiving application traffic. In a BIG-IP environment, a pool member is uniquely identified by the combination of its Node IP and its Service Port, and the Members tab is the primary interface for managing and auditing these specific member attributes.

The log message “Intercept exhaustion” indicates that the BIG-IP system has exhausted the available source port translations for one or more SNAT addresses. This occurs when too many concurrent client connections are being translated through a limited number of SNAT IP addresses, and all ephemeral source ports (typically ~64,000 per SNAT IP) are in use.

According to the BIG-IP Administration: Data Plane Configuration documentation:

Each SNAT IP address provides a finite number of available source ports.

When the number of concurrent connections exceeds the available port space, the BIG-IP logs an Intercept exhaustion error and new connections fail.

The recommended resolution is to increase the available SNAT resources by adding additional IP addresses to the SNAT pool.

Why the other options are incorrect:

A. Increase the timeout of the SNAT addressesIncreasing timeouts may actually worsen the problem by keeping ports allocated longer, accelerating port exhaustion.

B. Remove the SNAT pool and apply SNAT AutomapSNAT Automap uses the Self IP addresses on the egress VLAN, which may not provide additional capacity and can introduce routing or design issues. This is not a direct or recommended fix for SNAT exhaustion.

C. Remove an IP address from the SNAT poolThis would reduce the number of available source ports and further exacerbate the intercept exhaustion condition.

Correct Resolution:

By adding an IP address to the SNAT pool, the BIG-IP increases the total number of available source ports, alleviating intercept exhaustion and restoring successful client connections.

The requirement states that load balancing decisions must be based on the number of active connections on each pool member. This directly maps to the Least Connections (member) load balancing method.

According to the BIG-IP Administration: Data Plane Configuration documentation:

Least Connections (member) selects the pool member with the fewest active connections at the time of the request.

This method dynamically adapts to real-time traffic patterns and ensures that more heavily loaded pool members receive fewer new connections.

It is especially effective for web servers where connection duration may vary and equal distribution of active sessions is desired.

Why the other options are incorrect:

B. Round RobinDistributes connections sequentially without considering current load or active connections.

C. Ratio (member)Distributes traffic based on static ratios, not real-time connection counts.

D. Ratio (node)Uses predefined ratios at the node level and does not account for active connection counts.

Correct Resolution:

Using Least Connections (member) ensures that new connections are directed to the pool member currently handling the fewest active connections, meeting the stated requirement.

In the F5 BIG-IP ecosystem, a standard ICMP health monitor functions by sending an ICMP echo request to a target node and expecting an ICMP echo reply within a specified timeout period. When a node is marked " DOWN " despite the backend server being configured to respond to ICMP, the issue typically lies in the network path or the specific packet exchange between the BIG-IP ' s self IP and the node ' s IP. Running a tcpdump is the most effective next step because it provides a real-time packet capture of the actual monitor traffic leaving the BIG-IP and any return traffic coming back from the server. This allows the administrator to verify if the BIG-IP is actually sending the echo request, if the request is reaching the server, and if the server is indeed replying or if the reply is being dropped by an intermediate firewall or a security policy.

While other tools have their place, they are inappropriate for this specific layer 3/4 connectivity issue. A qkview is a comprehensive diagnostic file used primarily for F5 Support to analyze the entire system ' s state but is overkill for initial connectivity troubleshooting. An ssldump is used for inspecting SSL/TLS handshakes and encrypted payloads, which is irrelevant for a non-encrypted ICMP monitor. A curl command is a tool for testing HTTP/HTTPS application-level responses; it cannot be used to troubleshoot ICMP (ping) connectivity directly. By using tcpdump -ni < vlan_name > host < node_ip > , the administrator can see the ICMP " type 8 " (request) and " type 0 " (reply) packets, immediately identifying if the monitor failure is due to a " Destination Unreachable " message or a simple lack of response, thereby pinpointing the root cause in the data plane.

The log message " aggressive mode activated " indicates that the BIG-IP ' s adaptive connection management system (the " Sweeper " ) has detected that the system ' s memory or connection limits are reaching a critical threshold. To protect the system from crashing due to memory exhaustion (OOM), the BIG-IP enters Aggressive Mode, where it begins to proactively and rapidly reap (close) idle connections to free up resources for new incoming traffic.

To mitigate this and return the system to a healthy state, the administrator needs to reduce the overall resource footprint of existing connections. Decreasing the TCP profile Idle Timeout value (Option B) is the most effective administrative action. In many environments, the default idle timeout is 300 seconds (5 minutes). If a large number of connections remain " open " in the BIG-IP connection table long after the clients have stopped sending data, they consume valuable TMM (Traffic Management Microkernel) memory. By lowering the timeout (e.g., to 60 or 120 seconds), the BIG-IP can expire and remove these inactive entries much sooner, preventing the connection table from bloating and triggering the Sweeper ' s aggressive mode.

Conversely, increasing the timeout (Option C) would exacerbate the problem by keeping " dead " connections in memory even longer. Connection Mirroring (Option D) actually increases memory usage because every connection must be duplicated on the standby peer. An active-active configuration (Option A) might spread the load but does not address the underlying resource management issue on the individual units. Therefore, tightening the idle timers is the standard procedural fix for memory pressure caused by high connection volumes.

Application-specific health checks must validate application responses, which requires HTTP monitors with custom send/receive strings.

Forwarding (IP) virtual servers forward traffic based on routing decisions and do not use pools or pool members.

The pool member list in the exhibit shows servers listening on four distinct service ports: 80 (HTTP), 21 (FTP), 443 (HTTPS), and 22 (SSH). To accurately monitor a pool where members provide different services, the administrator must apply monitors that correspond to each of those specific services.

The key to this configuration is the Availability Requirement (also known as " Monitor Rule " ). If the administrator sets the requirement to " All health monitors, " every single monitor (HTTP, HTTPS, FTP, and SSH) must pass for a member to be marked " Up. " This would cause an immediate failure: for example, a server listening only on port 80 would fail the HTTPS, FTP, and SSH checks, resulting in the member being marked " Down " even if the HTTP service is healthy.

To ensure accuracy, the administrator should apply all four relevant monitors and set the Availability Requirement to at least one monitor. With this setting, the BIG-IP marks a member as " Up " if any of the assigned monitors return a successful response. Therefore, the member at 10.200.50.210:80 will stay " Up " as long as the HTTP monitor passes, even though it fails the FTP, HTTPS, and SSH monitors. This configuration allows a single pool to contain diverse service types while ensuring that the specific port defined for each member is verified correctly. Option C is incorrect because ICMP only checks if the IP is alive, not if the specific service port is functioning.

In the BIG-IP object hierarchy, a Pool Member is a child object of a Node . A pool member represents a specific IP:Port combination, while the parent node represents the underlying server IP address. When a node is disabled , all pool members that are children of that node are rendered unavailable — regardless of the individual pool member ' s own health or enabled state.

The Network Map depicted shows the pool member (192.168.212.30:80) with its Parent Node (192.168.212.30) disabled. This parent-child dependency means that even if the pool member itself is healthy and enabled, the disabled node cascades its unavailable state downward, causing the pool to have no available members and rendering VS_HTTP unresponsive to incoming connections.

The other options can be eliminated as follows:

Pool member monitor failed — the monitor status is not indicated as failed in the Network Map display.

Pool member is disabled — the pool member itself is not shown as disabled; the parent node is.

Node monitor failed — no monitor failure is indicated; the node ' s administrative state is explicitly disabled.

Understanding the node-to-member inheritance of availability state is fundamental to accurate BIG-IP traffic troubleshooting.

When a pool contains servers with heterogeneous hardware capabilities (differing CPU, RAM, or connection limits), a static load balancing method like Round Robin is ineffective because it distributes requests equally, regardless of the server ' s capacity. To optimize traffic distribution for newer, more powerful servers, a dynamic or weighted method is required.

Weighted Least Connections (member): This is the ideal method for this scenario. It combines two factors:

Least Connections: It first checks the current active connection count to ensure traffic goes to the least busy server.

Weight (Ratio): It allows the administrator to assign a " Ratio " value to each pool member. Newer servers can be assigned a higher ratio (e.g., 3) while older servers are assigned a lower ratio (e.g., 1). The BIG-IP system uses these weights to disproportionately favor the newer servers even when connection counts are similar.

Why other options are incorrect:

Global Availability: This is primarily a GSLB (Global Server Load Balancing) or specific LTM priority group concept where traffic is sent to the first available member in a list until it fails, then moves to the next. It does not load balance based on capacity.

Round Robin: This passes each new connection request to the next server in line, treating the old and new servers exactly the same.

Least Connections (member): While this sends traffic to the server with the fewest active connections, it assumes all servers are equal. If an old server and a new server both have 10 connections, they are treated as equally capable of taking the 11th, which is not true in this scenario.

With manual resume enabled, BIG-IP does not automatically return a pool member to service after recovery. The member remains offline until manually re-enabled.

F5 provides several pre-configured (built-in) TCP profiles designed to optimize traffic for different network conditions. When an administrator is looking to improve general performance but does not have a specific, narrow use case (like strictly mobile or strictly long-haul WAN), the f5-tcp-progressive profile is the recommended starting point.

The f5-tcp-progressive profile is designed as a modern, high-performance replacement for the older default TCP settings. It incorporates several advanced congestion control algorithms and buffer management techniques that allow it to adapt dynamically to varying network latencies and packet loss scenarios. Unlike f5-tcp-legacy (Option A), which uses older, less efficient stacks, the progressive profile leverages modern TMM (Traffic Management Microkernel) optimizations.

While f5-tcp-mobile (Option B) is highly effective for high-loss cellular networks and f5-tcp-wan (Option D) is optimized for high-latency long-distance links, they can sometimes be too aggressive or poorly suited for standard campus or data center environments. The f5-tcp-progressive profile acts as a " best-of-all-worlds " template that typically provides an immediate performance boost for most web applications by improving window scaling and fast-retransmit behavior. Therefore, it is the procedural " first step " in TCP performance tuning before moving to more specialized, niche profiles.

A red diamond indicates that the Virtual Server is enabled but unavailable, typically due to all pool members being down.

When a single IP address (Virtual Server) must host multiple hostnames (e.g., site1.com, site2.com) over HTTPS, the BIG-IP must be able to present the correct SSL certificate for each site during the initial handshake. This requirement is handled by a technology called Server Name Indication (SNI).

To implement this on the BIG-IP, the administrator must use a Client SSL profile because the encryption is occurring between the user (Client) and the Virtual Server. Within the Client SSL profile settings, there is a field called Server Name. By creating multiple Client SSL profiles—each with a different " Server Name " (the hostname) and its corresponding certificate—and attaching them all to the same Virtual Server, the BIG-IP can inspect the SNI extension in the client ' s " Client Hello " packet. It then selects the specific profile that matches the requested hostname.

Option A is incorrect because " Client Name " is not the standard field name used for SNI hostname matching in the profile. Options B and C refer to Server SSL profiles, which manage the connection between the BIG-IP and the backend servers. While Server SSL profiles also have a " Server Name " field (used to send SNI to a backend server), the requirement here is to provide encrypted communication for " users " accessing the virtual server, which definitively requires a Client-side SSL configuration.

SNAT Automap is a common configuration that replaces the client ' s original source IP address with one of the BIG-IP ' s self IP addresses. This ensures that the backend servers send return traffic back through the BIG-IP, which is necessary for the ADC to process the traffic correctly. However, a side effect of SNAT is that the backend servers only see the BIG-IP ' s IP in their logs, losing visibility into the true identity of the client.

To resolve this while still using SNAT for routing purposes, the administrator must configure the BIG-IP to " pass " the client ' s IP address at the application layer. This is achieved by using an HTTP Profile with the Insert X-Forwarded-For setting enabled. When this profile is applied to the Virtual Server, the BIG-IP intercepts the HTTP request, adds a header (X-Forwarded-For) containing the client ' s original IP, and then forwards the modified request to the server. The backend web server can then be configured to read this header and log the original client IP instead of the BIG-IP ' s SNAT address.

Other options are incorrect for this requirement. Performance (HTTP) (Option A) is a virtual server type optimized for speed but often lacks the full Layer 7 header manipulation capabilities of a Standard Virtual Server. SNAT Pool with the client IP (Option C) is technically impossible as SNAT pools use static, pre-defined IPs. There is no such thing as an HTTP Transparent profile (Option D) in standard BIG-IP administration for this purpose. The X-Forwarded-For header insertion within the HTTP profile is the standard procedural method for maintaining client visibility in SNAT-enabled environments.