NCP-GENL flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Response file install

An unattended installation method that supplies answers to installer prompts from a predefined file, enabling consistent, repeatable deployments without manual interaction.

Configuration precedence

The order in which settings are resolved when multiple sources define the same key, typically command line over environment over local file over global file over defaults.

Idempotent configuration

A property of an operation where applying it multiple times yields the same end state as applying it once, which is essential for safe automation and drift correction.

Package lock contention

A condition where two package manager processes attempt to modify the same database simultaneously; one waits or fails until the lock is released.

Hardening baseline

A documented set of minimum security settings applied after installation, covering services, permissions, authentication, and patch level before a system joins production.

Usable hosts formula

For an IPv4 subnet, usable host addresses equal two raised to the power of thirty-two minus the prefix length, minus two for network and broadcast addresses.

TCP handshake

The three-step connection setup where the client sends SYN, the server replies SYN-ACK, and the client sends ACK, establishing synchronized sequence numbers.

Longest prefix match

A routing rule stating that when multiple routes match a destination, the route with the most specific prefix, meaning the largest prefix length, is selected.

802.1Q VLAN tag

A four-byte field inserted into an Ethernet frame that identifies the VLAN membership, allowing a single trunk link to carry traffic for multiple VLANs.

Path MTU discovery

A technique that determines the smallest maximum transmission unit along a network path by setting the don't-fragment bit and processing fragmentation-needed ICMP messages.

SNAT versus DNAT

Source NAT rewrites the source address of outbound packets for masquerading, while destination NAT rewrites the destination address for inbound port forwarding.

RAID parity overhead

The capacity consumed by redundancy: RAID5 loses one disk worth, RAID6 loses two, and RAID10 loses half of raw capacity to mirroring.

Inode exhaustion

A condition where the file system runs out of inode structures despite free data blocks, preventing creation of new files, detectable with df -i.

LVM volume stack

The layered storage model where physical volumes are pooled into a volume group, from which logical volumes are allocated and then formatted with a file system.

Snapshot consistency

The guarantee that a point-in-time copy reflects a crash-consistent or application-consistent state, achieved by flushing buffers or quiescing the application before capture.

NFS root_squash

An export option that maps the remote root user to an unprivileged anonymous user, preventing a client root from gaining root privileges on the server.

SMART reallocated sectors

A disk attribute counting sectors remapped to spare areas; a rising value indicates surface degradation and predicts impending drive failure.

Sticky bit

A directory permission bit that restricts deletion of files to the file owner, the directory owner, or root, commonly used on shared temporary directories.

SELinux enforcing mode

A mandatory access control mode that denies access not explicitly allowed by policy and logs the denial, as opposed to permissive mode which only logs.

Multi-factor authentication

Authentication requiring two or more independent credential categories, such as a password plus a hardware token, reducing risk from a single compromised factor.

Certificate chain validation

The process of verifying that a server certificate is signed by a trusted authority, is unexpired, matches the hostname, and has not been revoked.

Password hashing

Storing credentials as a salted, deliberately slow one-way hash such as bcrypt or Argon2 so that stolen hashes are expensive to reverse by brute force.

USE method

A performance analysis approach that examines utilization, saturation, and errors for every system resource to locate bottlenecks systematically.

Golden signals

Four key service metrics for monitoring: latency, traffic, errors, and saturation, used to detect user-impacting problems before component-level causes.

Load average

The one, five, and fifteen minute averages of runnable plus uninterruptible processes; compared with core count, it reveals CPU queuing and saturation.

Memory pressure indicators

Signals such as low available memory, swap-in and swap-out activity, and out-of-memory killer messages that indicate the system is short on usable memory.

Disk await metric

The average time in milliseconds that I/O requests wait, including queue time; high await with high utilization indicates a saturated storage device.

Layered network troubleshooting

An ordered diagnostic approach that checks link, then IP reachability, then port listening, then application response, isolating faults at the lowest failing layer.

Shell strict mode

A script preamble using set -euo pipefail that aborts on command failure, unset variables, and pipeline errors, preventing silent partial execution.

Cron schedule fields

Five fields representing minute, hour, day of month, month, and day of week; an asterisk matches all values and a slash defines step intervals.

Runtime secret injection

Supplying credentials to a process at execution time from a vault or environment rather than storing them in source code or configuration files.

Dry-run validation

Executing an automation tool in a mode that reports intended changes without applying them, such as a plan or check flag, to catch errors safely.