Command Breakdown: Background Execution with Stream Redirection

The following command demonstrates several powerful Linux shell features combined:

nohup data-processor -f queries.hql > execution.log 2>&1 &

Component Analysis

1. nohup (No Hang Up)

The nohup utility makes a process immune to SIGHUP signals. When a terminal session terminates—whether through SSH disconnection, terminal closure, or logout—the operating system normally sends SIGHUP to all child processes, causing them to exit. By prefixing a command with nohup, the process ignores this signal and continues running independently of the controlling terminal.

2. data-processor -f queries.hql

This represents the target application and its arguments. The -f flag instructs the program to read input from a file—in this case, queries.hql, which presumably contains HiveQL statements for batch processing.

3. > execution.log

The single-angle bracket redirects standard output (stdout, file descriptor 1) from the command to execution.log. This captures all normal program output that would otherwise display in the terminal.

4. 2>&1

This construct merges standard error (stderr, file descriptor 2) with standard output. The &1 syntax references the current destination of stdout, ensuring both output streams are written to the same file. Without this, errors would still appear in the terminal or be lost when the session ends.

5. &

The trailing apmersand places the entire command pipeline into background execution. The shell immediately returns control, allowing continued interactive use while the process runs in a subshell. The combination with nohup ensures the background process survives terminal termination.

Practical Redirection Patterns

Stream Consolidation and Splitting

To display output on screen while simultaneously logging it:

myapp -c config.yaml 2>&1 | tee combined-output.log

For seperate log files without mixing streams:

myapp > stdout.log 2> stderr.log

Error Suppression and Filtering

To discard errors entirely while preserving standard output:

critical-task > results.txt 2>/dev/null

To capture only specific error patterns for alerting:

batch-job 2>&1 >/dev/null | grep -i "fatal\|critical" | alert-monitor

Advanced File Descriptor Manipulation

Swapping stdout and stderr using a temporary descriptor:

complex-operation 3>&1 1>&2 2>&3

Dynamically processing error streams with process substitution:

data-loader 2> >(error-parser --format=json > error-report.json)

Script-Wide Redirection

Using exec to apply redirection to all subsequent commands in a script:

#!/bin/bash
exec 1>>application.log
exec 2>>error-capture.log

process-data --input=dataset1.csv
validate-results --threshold=0.95
generate-report --type=summary

Debugging with Execution Tracing

Enable verbose mode to see each command and its redirection behavior:

set -x
pipeline-stage1 | pipeline-stage2 > final-output.log 2>&1
set +x

Conditional Logging Based on Exit Status

Capture output only when a command fails:

if ! deployment-script > deploy.log 2>&1; then
    cat deploy.log | mail -s "Deployment Failed" ops-team@example.com
fi