Add -j flag to send_dir.sh for concurrent, non-blocking uploads

This commit is contained in:
2026-07-02 15:06:59 -04:00
parent 871c682eb2
commit 32317537db

View File

@@ -13,6 +13,8 @@
# -u URL server base URL (default: http://127.0.0.1:8080, or $FS_URL) # -u URL server base URL (default: http://127.0.0.1:8080, or $FS_URL)
# -b batch: send all files in one multipart request # -b batch: send all files in one multipart request
# -r recurse into subdirectories # -r recurse into subdirectories
# -j N upload up to N files concurrently (default: 1). A slow big
# file no longer blocks the others; results are tallied at the end.
# #
# Examples: # Examples:
# bin/send_dir.sh data/samples # bin/send_dir.sh data/samples
@@ -22,19 +24,26 @@ set -euo pipefail
url="${FS_URL:-http://127.0.0.1:8080}" url="${FS_URL:-http://127.0.0.1:8080}"
batch=0 batch=0
recurse=0 recurse=0
jobs=1
while getopts ":u:brh" opt; do while getopts ":u:brj:h" opt; do
case "$opt" in case "$opt" in
u) url="$OPTARG" ;; u) url="$OPTARG" ;;
b) batch=1 ;; b) batch=1 ;;
r) recurse=1 ;; r) recurse=1 ;;
h) sed -n '2,20p' "$0"; exit 0 ;; j) jobs="$OPTARG" ;;
h) sed -n '2,22p' "$0"; exit 0 ;;
\?) echo "unknown option: -$OPTARG" >&2; exit 2 ;; \?) echo "unknown option: -$OPTARG" >&2; exit 2 ;;
:) echo "option -$OPTARG needs an argument" >&2; exit 2 ;; :) echo "option -$OPTARG needs an argument" >&2; exit 2 ;;
esac esac
done done
shift $((OPTIND - 1)) shift $((OPTIND - 1))
if ! [[ "$jobs" =~ ^[0-9]+$ ]] || [[ "$jobs" -lt 1 ]]; then
echo "-j needs a positive integer, got: $jobs" >&2
exit 2
fi
dir="${1:-}" dir="${1:-}"
if [[ -z "$dir" ]]; then if [[ -z "$dir" ]]; then
echo "usage: $0 [-u URL] [-b] [-r] DIR" >&2 echo "usage: $0 [-u URL] [-b] [-r] DIR" >&2
@@ -77,14 +86,33 @@ if [[ "$batch" -eq 1 ]]; then
echo "-> HTTP $code" echo "-> HTTP $code"
[[ "$code" == "202" ]] [[ "$code" == "202" ]]
else else
# One request per file. # One request per file, up to $jobs in flight at once. Each upload runs in the
ok=0 # background so a slow (big) file doesn't stall the ones behind it. Every job
fail=0 # writes its HTTP status to a per-file line in $results; we tally after all
for f in "${files[@]}"; do # jobs drain.
results="$(mktemp)"
trap 'rm -f "$results"' EXIT
# Upload one file: print progress and append "code<TAB>path" to $results.
upload_one() {
local f="$1" code
code=$(curl -sS -o /dev/stderr -w '%{http_code}' -F "file=@${f}" "$endpoint") code=$(curl -sS -o /dev/stderr -w '%{http_code}' -F "file=@${f}" "$endpoint")
echo " -> HTTP $code ${f}" echo " -> HTTP $code ${f}"
if [[ "$code" == "202" ]]; then ok=$((ok + 1)); else fail=$((fail + 1)); fi printf '%s\t%s\n' "$code" "$f" >> "$results"
}
for f in "${files[@]}"; do
# Throttle: wait for a slot to free up before launching the next upload.
# (|| true: a failed upload's exit status must not trip `set -e` here.)
while (( $(jobs -r -p | wc -l) >= jobs )); do
wait -n || true
done
upload_one "$f" &
done done
wait || true
ok=$(grep -c $'^202\t' "$results" || true)
fail=$(( ${#files[@]} - ok ))
echo "done: $ok accepted, $fail failed" echo "done: $ok accepted, $fail failed"
[[ "$fail" -eq 0 ]] [[ "$fail" -eq 0 ]]
fi fi