test: check dedup of large files (requiring chunked) + test checunk upload via API

This commit is contained in:
Edouard Vanbelle
2026-05-13 16:20:58 +02:00
parent 28e25f9d16
commit 739235faf0
4 changed files with 395 additions and 1 deletions
+7 -1
View File
@@ -86,7 +86,13 @@ log "Server is ready."
log "Running Hurl tests..."
for T in "$WEBDAV_DIR"/test_*.sh; do
bash "$T"
if bash "$T"
then
echo $'\e[32m'"Success $T"$'\e[0m' >&2
else
echo $'\e[31m'"Failure $T"$'\e[0m' >&2
false
fi
done
log "All tests passed."
+174
View File
@@ -0,0 +1,174 @@
#!/usr/bin/env bash
# =============================================================
# OxiCloud – Chunked upload API + dedup check
# =============================================================
# Uploads free_video_over_1MB.mp4 (2760653 bytes) in 3 chunks
# of 1 MiB each via the TUS-like chunked upload API, then:
# 1. Verifies the file appears in folder listing with video/mp4 MIME type
# 2. Checks GET /api/dedup/check/{hash} → ref_count == 1
#
# BLAKE3 hash of free_video_over_1MB.mp4:
# 95d42b25a2d39f24f1b2f38bf1b947d4ec74201271a98ea0e76a9cea421eff80
#
# Prerequisites:
# - Server running at base_url with credentials from test.env
# - OXICLOUD_ENABLE_AUTH=true
# - jq, dd in PATH
#
# Run (from repo root):
# bash tests/webdav/test_chunked_upload_dedup.sh
# =============================================================
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
REPO_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
cd "$SCRIPT_DIR"
source test.env
source common.sh
# ── helpers ──────────────────────────────────────────────────────────────────
PASS=0
FAIL=0
pass() { PASS=$(( PASS + 1 )); echo " PASS: $*"; }
fail() { FAIL=$(( FAIL + 1 )); echo " FAIL: $*" >&2; exit 1; }
rest_get() { curl -s -H "Authorization: Bearer $TOKEN" "$base_url$1"; }
rest_delete() { curl -s -o /dev/null -w "%{http_code}" -X DELETE -H "Authorization: Bearer $TOKEN" "$base_url$1"; }
dedup_check() { curl -s -H "Authorization: Bearer $TOKEN" "$base_url/api/dedup/check/$1"; }
purge_from_trash() {
local name="$1"
local tid
tid=$(rest_get "/api/trash" \
| jq -r --arg n "$name" 'first(.[] | select(.name == $n) | .id) // empty')
[[ -n "$tid" ]] && rest_delete "/api/trash/$tid" > /dev/null || true
}
# ── fixture ───────────────────────────────────────────────────────────────────
BLOB_HASH="95d42b25a2d39f24f1b2f38bf1b947d4ec74201271a98ea0e76a9cea421eff80"
FIXTURE="$REPO_ROOT/tests/fixtures/free_video_over_1MB.mp4"
[[ -f "$FIXTURE" ]] || { echo "Missing fixture: $FIXTURE" >&2; exit 1; }
REMOTE_NAME="chunked-upload-test.mp4"
FILE_SIZE=2760653
CHUNK_SIZE=1048576 # 1 MiB — minimum accepted by the server
TOTAL_CHUNKS=3 # ceil(2760653 / 1048576) = 3
echo
echo "=== Chunked upload API + dedup check ==="
echo
# ── authenticate ──────────────────────────────────────────────────────────────
oxicloud_login
# ── home folder ───────────────────────────────────────────────────────────────
HOME_FOLDER_ID=$(rest_get "/api/folders" | jq -r '.[0].id')
[[ -n "$HOME_FOLDER_ID" && "$HOME_FOLDER_ID" != "null" ]] \
|| fail "Could not retrieve home folder ID"
echo " home folder id: $HOME_FOLDER_ID"
# ── idempotent pre-test cleanup ───────────────────────────────────────────────
EXISTING_ID=$(rest_get "/api/files?folder_id=$HOME_FOLDER_ID" \
| jq -r --arg n "$REMOTE_NAME" 'first(.[] | select(.name == $n) | .id) // empty')
if [[ -n "$EXISTING_ID" ]]; then
echo " cleanup: deleting existing $REMOTE_NAME (id=$EXISTING_ID)"
rest_delete "/api/files/$EXISTING_ID" > /dev/null
fi
purge_from_trash "$REMOTE_NAME"
# ── Step 1: Create upload session ─────────────────────────────────────────────
echo " step 1: POST /api/uploads (create session, $TOTAL_CHUNKS chunks of ${CHUNK_SIZE}B)..."
SESSION=$(curl -s \
-X POST \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d "{\"filename\":\"$REMOTE_NAME\",\"folder_id\":\"$HOME_FOLDER_ID\",\"content_type\":\"video/mp4\",\"total_size\":$FILE_SIZE,\"chunk_size\":$CHUNK_SIZE}" \
"$base_url/api/uploads")
UPLOAD_ID=$(jq -r '.upload_id' <<< "$SESSION")
[[ -n "$UPLOAD_ID" && "$UPLOAD_ID" != "null" ]] \
|| fail "POST /api/uploads: could not get upload_id (response: $SESSION)"
pass "Upload session created: upload_id=$UPLOAD_ID"
# ── Step 2: Upload chunks ─────────────────────────────────────────────────────
echo " step 2: PATCH chunks 0..$(( TOTAL_CHUNKS - 1 ))..."
for (( i=0; i<TOTAL_CHUNKS; i++ )); do
PATCH_RESP=$(dd if="$FIXTURE" bs="$CHUNK_SIZE" skip="$i" count=1 2>/dev/null \
| curl -s \
-X PATCH \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/octet-stream" \
--data-binary @- \
"$base_url/api/uploads/$UPLOAD_ID?chunk_index=$i")
IS_COMPLETE=$(jq -r '.is_complete' <<< "$PATCH_RESP")
BYTES=$(jq -r '.bytes_received' <<< "$PATCH_RESP")
[[ -n "$BYTES" && "$BYTES" != "null" ]] \
|| fail "PATCH chunk $i: unexpected response: $PATCH_RESP"
echo " chunk $i: bytes_received=$BYTES is_complete=$IS_COMPLETE"
done
pass "All $TOTAL_CHUNKS chunks uploaded"
# ── Step 3: Complete the upload ───────────────────────────────────────────────
echo " step 3: POST /api/uploads/$UPLOAD_ID/complete..."
COMPLETE_RESP=$(curl -s \
-X POST \
-H "Authorization: Bearer $TOKEN" \
"$base_url/api/uploads/$UPLOAD_ID/complete")
FILE_ID=$(jq -r '.file_id' <<< "$COMPLETE_RESP")
[[ -n "$FILE_ID" && "$FILE_ID" != "null" ]] \
|| fail "POST complete: could not get file_id (response: $COMPLETE_RESP)"
pass "Upload complete: file_id=$FILE_ID"
# ── Step 4: Verify file appears in folder listing with correct MIME type ──────
echo " step 4: verify file in folder listing with video/mp4 MIME type..."
LISTING=$(rest_get "/api/files?folder_id=$HOME_FOLDER_ID")
LISTED_FILE=$(jq -r --arg id "$FILE_ID" 'first(.[] | select(.id == $id))' <<< "$LISTING")
[[ -n "$LISTED_FILE" && "$LISTED_FILE" != "null" ]] \
|| fail "File $FILE_ID not found in folder listing"
MIME=$(jq -r '.mime_type' <<< "$LISTED_FILE")
[[ "$MIME" == "video/mp4" ]] \
|| fail "Expected MIME type video/mp4, got: $MIME"
pass "File listed with MIME type: $MIME"
# ── Step 5: Dedup check → ref_count == 1 ─────────────────────────────────────
echo " step 5: GET /api/dedup/check/$BLOB_HASH..."
RESP=$(dedup_check "$BLOB_HASH")
EXISTS=$(jq -r '.exists' <<< "$RESP")
RC=$( jq -r '.ref_count' <<< "$RESP")
[[ "$EXISTS" == "true" ]] \
|| fail "dedup/check: expected exists=true, got $EXISTS (response: $RESP)"
[[ "$RC" == "1" ]] \
|| fail "dedup/check: expected ref_count=1, got $RC"
pass "ref_count == 1: blob registered after chunked upload"
# ── cleanup ───────────────────────────────────────────────────────────────────
echo " cleanup..."
ST=$(rest_delete "/api/files/$FILE_ID")
[[ "$ST" == "204" ]] || fail "DELETE file expected 204, got $ST"
purge_from_trash "$REMOTE_NAME"
pass "Cleanup complete"
# ── summary ───────────────────────────────────────────────────────────────────
echo
echo "Results: $PASS passed, $FAIL failed."
[[ "$FAIL" -eq 0 ]] && echo "All tests passed." || exit 1
+214
View File
@@ -0,0 +1,214 @@
#!/usr/bin/env bash
# =============================================================
# OxiCloud – Dedup ref_count: multi-chunk CDC file via WebDAV
# =============================================================
# Case A (files < 64kB = 1 chunk) is treated in tests/api/dedup_blob_cleanup.hurl
# Validates Case B of the cleanup_if_orphaned fix:
# file_hash ≠ chunk_hashes (file split into 8 CDC chunks)
#
# free_video_over_1MB.mp4 — 2.6 MB, 8 CDC chunks
# BLAKE3: 95d42b25a2d39f24f1b2f38bf1b947d4ec74201271a98ea0e76a9cea421eff80
#
# Sequence:
# 1. PUT video as file A → new manifest (ref_count=1, 8 chunk blobs)
# 2. PUT video as file B → dedup hit (ref_count=2, chunks unchanged)
# 3. /api/dedup/check → ref_count == 2
# 4. Permanently delete file A
# 5. /api/dedup/check → ref_count == 1 (chunks must NOT be freed)
# 6. Permanently delete file B
# 7. /api/dedup/check → exists == false (manifest + chunks cleaned up)
#
# Case B regression: cleanup_if_orphaned must decrement chunk_manifests
# ref_count without touching chunk blobs (the PG trigger is a no-op for
# multi-chunk files because file_hash is not stored in storage.blobs).
#
# Prerequisites:
# - Server running at base_url with credentials from test.env
# - OXICLOUD_ENABLE_AUTH=true
# - jq in PATH
#
# Run (from repo root):
# bash tests/webdav/test_dedup_webdav_multichunk.sh
# =============================================================
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
REPO_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
cd "$SCRIPT_DIR"
source test.env
source common.sh
# ── helpers ──────────────────────────────────────────────────────────────────
PASS=0
FAIL=0
pass() { PASS=$(( PASS + 1 )); echo " PASS: $*"; }
fail() { FAIL=$(( FAIL + 1 )); echo " FAIL: $*" >&2; exit 1; }
webdav_put() {
local remote_name="$1" local_file="$2" mime="${3:-application/octet-stream}"
curl -s -o /dev/null -w "%{http_code}" \
-X PUT \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: $mime" \
--data-binary "@$local_file" \
"$base_url/webdav/$remote_name"
}
webdav_delete() {
local remote_name="$1"
curl -s -o /dev/null -w "%{http_code}" \
-X DELETE \
-H "Authorization: Bearer $TOKEN" \
"$base_url/webdav/$remote_name"
}
rest_get() { curl -s -H "Authorization: Bearer $TOKEN" "$base_url$1"; }
rest_delete() { curl -s -o /dev/null -w "%{http_code}" -X DELETE -H "Authorization: Bearer $TOKEN" "$base_url$1"; }
dedup_check() { curl -s -H "Authorization: Bearer $TOKEN" "$base_url/api/dedup/check/$1"; }
purge_from_trash() {
local name="$1"
local tid
tid=$(rest_get "/api/trash" \
| jq -r --arg n "$name" 'first(.[] | select(.name == $n) | .id) // empty')
[[ -n "$tid" ]] && rest_delete "/api/trash/$tid" > /dev/null || true
}
# ── fixture ───────────────────────────────────────────────────────────────────
# free_video_over_1MB.mp4 → 2.6 MB → 8 CDC chunks → confirmed Case B
BLOB_HASH="95d42b25a2d39f24f1b2f38bf1b947d4ec74201271a98ea0e76a9cea421eff80"
FIXTURE="$REPO_ROOT/tests/fixtures/free_video_over_1MB.mp4"
[[ -f "$FIXTURE" ]] || { echo "Missing fixture: $FIXTURE" >&2; exit 1; }
FILE_A="webdav-dedup-mc-a.mp4"
FILE_B="webdav-dedup-mc-b.mp4"
echo
echo "=== Dedup ref_count: multi-chunk CDC (Case B) via WebDAV ==="
echo
# ── authenticate ──────────────────────────────────────────────────────────────
oxicloud_login
# ── home folder ───────────────────────────────────────────────────────────────
HOME_FOLDER_ID=$(rest_get "/api/folders" | jq -r '.[0].id')
[[ -n "$HOME_FOLDER_ID" && "$HOME_FOLDER_ID" != "null" ]] \
|| fail "Could not retrieve home folder ID"
echo " home folder id: $HOME_FOLDER_ID"
# ── idempotent pre-test cleanup ───────────────────────────────────────────────
for REMOTE in "$FILE_A" "$FILE_B"; do
EXISTING_ID=$(rest_get "/api/files?folder_id=$HOME_FOLDER_ID" \
| jq -r --arg n "$REMOTE" 'first(.[] | select(.name == $n) | .id) // empty')
if [[ -n "$EXISTING_ID" ]]; then
echo " cleanup: deleting existing $REMOTE (id=$EXISTING_ID)"
rest_delete "/api/files/$EXISTING_ID" > /dev/null
fi
purge_from_trash "$REMOTE"
done
# ── Step 1: Upload file A ─────────────────────────────────────────────────────
echo " step 1: PUT $FILE_A..."
STATUS=$(webdav_put "$FILE_A" "$FIXTURE" "video/mp4")
[[ "$STATUS" == "204" ]] || fail "PUT $FILE_A expected 204, got $STATUS"
pass "PUT $FILE_A → 204 (new manifest, 8 chunk blobs created)"
# ── Step 2: Upload file B (same content, different name → dedup hit) ──────────
echo " step 2: PUT $FILE_B (same bytes → dedup hit)..."
STATUS=$(webdav_put "$FILE_B" "$FIXTURE" "video/mp4")
[[ "$STATUS" == "204" ]] || fail "PUT $FILE_B expected 204, got $STATUS"
pass "PUT $FILE_B → 204 (dedup hit: manifest ref_count → 2, chunks unchanged)"
# ── Resolve file IDs ──────────────────────────────────────────────────────────
LISTING=$(rest_get "/api/files?folder_id=$HOME_FOLDER_ID")
FILE_A_ID=$(jq -r --arg n "$FILE_A" '.[] | select(.name == $n) | .id' <<< "$LISTING")
FILE_B_ID=$(jq -r --arg n "$FILE_B" '.[] | select(.name == $n) | .id' <<< "$LISTING")
[[ -n "$FILE_A_ID" && "$FILE_A_ID" != "null" ]] || fail "File A not found in listing"
[[ -n "$FILE_B_ID" && "$FILE_B_ID" != "null" ]] || fail "File B not found in listing"
[[ "$FILE_A_ID" != "$FILE_B_ID" ]] \
|| fail "File A and B share the same ID — dedup must create two distinct records"
pass "Two distinct file records: A=$FILE_A_ID B=$FILE_B_ID"
# ── Step 3: ref_count == 2 ────────────────────────────────────────────────────
echo " step 3: dedup/check → expect ref_count=2..."
RESP=$(dedup_check "$BLOB_HASH")
EXISTS=$(jq -r '.exists' <<< "$RESP")
RC=$( jq -r '.ref_count' <<< "$RESP")
[[ "$EXISTS" == "true" ]] \
|| fail "dedup/check: expected exists=true, got $EXISTS (response: $RESP)"
[[ "$RC" == "2" ]] \
|| fail "dedup/check: expected ref_count=2, got $RC"
pass "ref_count == 2: both files reference the same 8-chunk blob"
# ── Step 4: Permanently delete file A ────────────────────────────────────────
# For multi-chunk files: PG trigger is a no-op (file_hash not in storage.blobs).
# cleanup_if_orphaned must decrement chunk_manifests.ref_count only (2→1)
# and leave the 8 chunk blobs untouched.
echo " step 4: trash + permanently delete $FILE_A..."
ST=$(rest_delete "/api/files/$FILE_A_ID")
[[ "$ST" == "204" ]] || fail "DELETE $FILE_A expected 204, got $ST"
TRASH_A=$(rest_get "/api/trash" \
| jq -r --arg n "$FILE_A" 'first(.[] | select(.name == $n) | .id) // empty')
[[ -n "$TRASH_A" ]] || fail "File A not found in trash"
ST=$(rest_delete "/api/trash/$TRASH_A")
[[ "$ST" == "200" ]] || fail "Permanent delete file A expected 200, got $ST"
pass "File A permanently deleted"
# ── Step 5: ref_count == 1 — chunk blobs must still be alive ─────────────────
echo " step 5: dedup/check → expect ref_count=1 (chunks must survive)..."
RESP=$(dedup_check "$BLOB_HASH")
EXISTS=$(jq -r '.exists' <<< "$RESP")
RC=$( jq -r '.ref_count' <<< "$RESP")
[[ "$EXISTS" == "true" ]] \
|| fail "dedup/check: expected exists=true (file B still references blob), got $EXISTS"
[[ "$RC" == "1" ]] \
|| fail "dedup/check: expected ref_count=1, got $RC (chunk blobs may have been freed prematurely)"
pass "ref_count == 1: manifest decremented, all 8 chunk blobs still alive"
# ── Step 6: Permanently delete file B ────────────────────────────────────────
# ref_count hits 0 → manifest deleted, all 8 chunk blobs freed.
echo " step 6: trash + permanently delete $FILE_B..."
ST=$(rest_delete "/api/files/$FILE_B_ID")
[[ "$ST" == "204" ]] || fail "DELETE $FILE_B expected 204, got $ST"
TRASH_B=$(rest_get "/api/trash" \
| jq -r --arg n "$FILE_B" 'first(.[] | select(.name == $n) | .id) // empty')
[[ -n "$TRASH_B" ]] || fail "File B not found in trash"
ST=$(rest_delete "/api/trash/$TRASH_B")
[[ "$ST" == "200" ]] || fail "Permanent delete file B expected 200, got $ST"
pass "File B permanently deleted"
# ── Step 7: blob gone ─────────────────────────────────────────────────────────
echo " step 7: dedup/check → expect exists=false (manifest + chunks freed)..."
RESP=$(dedup_check "$BLOB_HASH")
EXISTS=$(jq -r '.exists' <<< "$RESP")
[[ "$EXISTS" == "false" ]] \
|| fail "dedup/check: expected exists=false after both files deleted, got $EXISTS"
pass "exists == false: manifest and all 8 chunk blobs cleaned up"
# ── summary ───────────────────────────────────────────────────────────────────
echo
echo "Results: $PASS passed, $FAIL failed."
[[ "$FAIL" -eq 0 ]] && echo "All tests passed." || exit 1