Curl Download Retry Shell Script Linux

CI mirrors and ISO pulls break mid-stream — a curl download retry shell script with resume (-C -) and bounded retries is what people bookmark instead of retyping flags.

What this script does

Downloads URL to output path (creates parent dirs)
Retries up to N times with sleep between attempts
Uses curl -L -f -C - for redirect follow and resume
Configurable connect and max time
Logs each attempt to stderr and optional log file

Prerequisites

curl installed
Disk space for output file
Outbound HTTPS allowed

Step 1: Save the script

sudo nano /usr/local/bin/download-retry.sh
sudo chmod +x /usr/local/bin/download-retry.sh

Step 2: Full script (scroll to read)

download-retry.sh

#!/usr/bin/env bash
set -euo pipefail

URL="${1:?usage: download-retry.sh URL OUTPUT_PATH}"
OUT="${2:?usage: download-retry.sh URL OUTPUT_PATH}"
RETRIES="${RETRIES:-5}"
WAIT="${WAIT_SEC:-10}"
CONNECT_TIMEOUT="${CONNECT_TIMEOUT:-30}"
MAX_TIME="${MAX_TIME:-0}"
LOG="${LOG_FILE:-/var/log/download-retry.log}"

log(){ echo "[$(date '+%F %T')] $*" | tee -a "$LOG"; }

mkdir -p "$(dirname "$OUT")"

attempt=1
while (( attempt <= RETRIES )); do
  log "Attempt $attempt/$RETRIES: $URL -> $OUT"
  extra=()
  [[ "$MAX_TIME" != "0" ]] && extra+=("--max-time" "$MAX_TIME")
  if curl -L -f -C - --connect-timeout "$CONNECT_TIMEOUT" "${extra[@]}" -o "$OUT" "$URL"; then
    log "OK downloaded $(wc -c <"$OUT") bytes"
    exit 0
  fi
  log "FAIL curl exit $?"
  (( attempt++ ))
  sleep "$WAIT"
done

log "Giving up after $RETRIES attempts"
exit 1

Scroll inside the box to read the full script.

Step 3: Configure settings

Verify with sha256sum after download for critical artifacts
Compare API health check for small payloads
Use credentials via netrc, not hard-coded passwords

Curl download script with retry and resume

Step 4: Test manually

./download-retry.sh https://example.com/index.html /tmp/example.html

RETRIES=5 ./download-retry.sh https://releases.ubuntu.com/22.04/ubuntu-22.04.5-desktop-amd64.iso.iso /tmp/test.iso

Schedule with cron

sudo crontab -e

Add:

30 4 * * * /usr/local/bin/download-retry.sh https://example.com/db-dump.sql.gz /backups/incoming/db.sql.gz >> /var/log/download-retry.log 2>&1

Curl Download File with Retry Shell Script Linux