# Linux Host and Filesystem Fundamentals ### Clarifying Questions to Ask - Which Linux distribution and shell should be assumed? - Are privileged commands allowed, or should the answer work as an ordinary user? - Is the system local, containerized, or running under an orchestration platform? ### Part 1: Inspect Memory and Explain Swap How would you inspect memory usage on a Linux host? Explain what swap is, why it exists, and why low “free” memory or nonzero swap use is not automatically evidence of a problem. **Candidate hint:** Separate process memory, reclaimable cache, memory pressure, and historical swap placement. #### What This Part Should Cover - Useful commands and fields rather than one unexplained number - Available memory, cache, swap, and pressure interpretation - Limits of a single snapshot ### Part 2: Working Directory and Listing Without `ls` Which command prints the current working directory? Give at least two safe ways to enumerate entries in the current directory without using `ls`, including hidden entries where practical. **Candidate hint:** Be explicit about recursion, dotfiles, quoting, and unusual filenames. #### What This Part Should Cover - `pwd` semantics - Shell expansion or `find`-based alternatives and their caveats - Filename-safe output when results feed another command ### Part 3: Threads and Processes Compare a Linux process with a thread. Discuss address spaces, resources, scheduling, communication, isolation, and failure implications. **Candidate hint:** Avoid saying that threads are not scheduled entities; focus on what is shared and what is private. #### What This Part Should Cover - Shared versus isolated state - Context, communication, synchronization, and fault boundaries - Practical performance trade-offs without absolute claims ### Part 4: Symbolic and Hard Links Explain the difference between a symbolic link and a hard link. Cover inode relationships, filesystem boundaries, directories, deletion of the target name, and how common tools display them. **Candidate hint:** Ask what directory entry each link points to and what keeps file data alive. #### What This Part Should Cover - Path indirection versus another name for one inode - Lifetime and broken-link behavior - Cross-filesystem and directory restrictions ### Part 5: Delete a File Named `-rf` A regular file in the current directory is literally named `-rf`. How can you delete that exact file without it being interpreted as command options? Explain why the command is safe. **Candidate hint:** Either terminate option parsing or make the operand stop beginning with a hyphen. #### What This Part Should Cover - A correctly quoted, exact operand - `--` or an explicit relative path - Awareness that quoting alone does not change a leading hyphen into a non-option ### Part 6: Recover After an Accidental Recursive Delete Suppose a recursive removal deleted the current directory's contents. Is there a general shell command that restores those files? Explain what deletion changes and give a responsible immediate response and recovery order. **Candidate hint:** Distinguish unlinking names from securely erasing blocks, but do not promise that unallocated data is recoverable. #### What This Part Should Cover - No universal undo and why continued writes reduce recovery chances - Backups, snapshots, version control, and still-open file descriptors - Filesystem-specific recovery as a last resort, not a guarantee ### What a Strong Answer Covers - Commands paired with an explanation of what they measure or modify - Safe handling of filenames and destructive operations - Kernel and filesystem concepts connected to practical troubleshooting - Appropriate caveats across distributions, filesystems, and runtime environments ### Follow-up Questions 1. Why can a process be killed by a memory limit while the host still reports available memory? 2. How would you pass arbitrary filenames safely from `find` to another command? 3. Why can an open file continue consuming disk after its pathname is deleted? 4. How do bind mounts differ from symbolic links?