qa/rgw: test that multipart re-upload does not leave any orphans

Runs a boto script that reuploads one part multiple times before
completing and then we check for any orphans.

Original boto script contributed by Matt Benjamin
<mbenjami@redhat.com> on top of which modifications were made.

Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>
(cherry picked from commit aeffd1b5986ad13f7a1f962cbff2fc72a5b21abe)

diff --git a/qa/suites/rgw/verify/tasks/mp_reupload.yaml b/qa/suites/rgw/verify/tasks/mp_reupload.yaml
new file mode 100644 (file)
index 0000000..d817a1c
--- /dev/null
+++ b/qa/suites/rgw/verify/tasks/mp_reupload.yaml
@@ -0,0 +1,5 @@
+tasks:
+- workunit:
+    clients:
+      client.0:
+        - rgw/test_rgw_s3_mp_reupload.sh

diff --git a/qa/workunits/rgw/test_rgw_s3_mp_reupload.py b/qa/workunits/rgw/test_rgw_s3_mp_reupload.py
new file mode 100755 (executable)
index 0000000..b3cb2d5
--- /dev/null
+++ b/qa/workunits/rgw/test_rgw_s3_mp_reupload.py
@@ -0,0 +1,121 @@
+import boto3
+import botocore.exceptions
+import sys
+import os
+import subprocess
+
+#boto3.set_stream_logger(name='botocore')
+
+# handles two optional system arguments:
+#   <bucket-name> : default is "bkt134"
+#   <0 or 1>      : 0 -> upload aborted, 1 -> completed; default is completed
+
+if len(sys.argv) >= 2:
+    bucket_name = sys.argv[1]
+else:
+    bucket_name = "bkt314738362229"
+print("bucket nams is %s" % bucket_name)
+
+complete_mpu = True
+if len(sys.argv) >= 3:
+    complete_mpu = int(sys.argv[2]) > 0
+
+versioned_bucket = False
+if len(sys.argv) >= 4:
+    versioned_bucket = int(sys.argv[3]) > 0
+
+rgw_host = os.environ['RGW_HOST']
+access_key = os.environ['RGW_ACCESS_KEY']
+secret_key = os.environ['RGW_SECRET_KEY']
+
+try:
+    endpoint='http://%s:%d' % (rgw_host, 80)
+    client = boto3.client('s3',
+                          endpoint_url=endpoint,
+                          aws_access_key_id=access_key,
+                          aws_secret_access_key=secret_key)
+    res = client.create_bucket(Bucket=bucket_name)
+except botocore.exceptions.EndpointConnectionError:
+    try:
+        endpoint='https://%s:%d' % (rgw_host, 443)
+        client = boto3.client('s3',
+                              endpoint_url=endpoint,
+                              verify=False,
+                              aws_access_key_id=access_key,
+                              aws_secret_access_key=secret_key)
+        res = client.create_bucket(Bucket=bucket_name)
+    except botocore.exceptions.EndpointConnectionError:
+        endpoint='http://%s:%d' % (rgw_host, 8000)
+        client = boto3.client('s3',
+                              endpoint_url=endpoint,
+                              aws_access_key_id=access_key,
+                              aws_secret_access_key=secret_key)
+        res = client.create_bucket(Bucket=bucket_name)
+
+print("endpoint is %s" % endpoint)
+
+if versioned_bucket:
+    res = client.put_bucket_versioning(
+        Bucket=bucket_name,
+        VersioningConfiguration={
+        'MFADelete': 'Disabled',
+        'Status': 'Enabled'}
+        )
+
+key = "mpu_test4"
+nparts = 2
+ndups = 11
+do_reupload = True
+
+part_path = "/tmp/mp_part_5m"
+subprocess.run(["dd", "if=/dev/urandom", "of=" + part_path, "bs=1M", "count=5"], check=True)
+
+f = open(part_path, 'rb')
+
+res = client.create_multipart_upload(Bucket=bucket_name, Key=key)
+mpu_id = res["UploadId"]
+
+print("start UploadId=%s" % (mpu_id))
+
+parts = []
+parts2 = []
+
+for ix in range(0,nparts):
+    part_num = ix + 1
+    f.seek(0)
+    res = client.upload_part(Body=f, Bucket=bucket_name, Key=key,
+                             UploadId=mpu_id, PartNumber=part_num)
+    # save
+    etag = res['ETag']
+    part = {'ETag': etag, 'PartNumber': part_num}
+    print("phase 1 uploaded part %s" % part)
+    parts.append(part)
+
+if do_reupload:
+    # just re-upload part 1
+    part_num = 1
+    for ix in range(0,ndups):
+        f.seek(0)
+        res = client.upload_part(Body=f, Bucket=bucket_name, Key=key,
+                                UploadId=mpu_id, PartNumber=part_num)
+        etag = res['ETag']
+        part = {'ETag': etag, 'PartNumber': part_num}
+        print ("phase 2 uploaded part %s" % part)
+
+        # save
+        etag = res['ETag']
+        part = {'ETag': etag, 'PartNumber': part_num}
+        parts2.append(part)
+
+if complete_mpu:
+    print("completing multipart upload, parts=%s" % parts)
+    res = client.complete_multipart_upload(
+        Bucket=bucket_name, Key=key, UploadId=mpu_id,
+        MultipartUpload={'Parts': parts})
+else:
+    print("aborting multipart upload, parts=%s" % parts)
+    res = client.abort_multipart_upload(
+        Bucket=bucket_name, Key=key, UploadId=mpu_id)
+
+# clean up
+subprocess.run(["rm", "-f", part_path], check=True)

diff --git a/qa/workunits/rgw/test_rgw_s3_mp_reupload.sh b/qa/workunits/rgw/test_rgw_s3_mp_reupload.sh
new file mode 100755 (executable)
index 0000000..5d73fd0
--- /dev/null
+++ b/qa/workunits/rgw/test_rgw_s3_mp_reupload.sh
@@ -0,0 +1,110 @@
+#!/usr/bin/env bash
+
+# INITIALIZATION
+
+mydir=$(dirname $0)
+data_pool=default.rgw.buckets.data
+orphan_list_out=/tmp/orphan_list.$$
+radoslist_out=/tmp/radoslist.$$
+rados_ls_out=/tmp/rados_ls.$$
+diff_out=/tmp/diff.$$
+
+rgw_host="$(hostname --fqdn)"
+echo "INFO: fully qualified domain name: $rgw_host"
+
+export RGW_ACCESS_KEY="0555b35654ad1656d804"
+export RGW_SECRET_KEY="h7GhxuBLTrlhVUyxSPUKUV8r/2EI4ngqJxD7iBdBYLhwluN30JaT3Q=="
+export RGW_HOST="${RGW_HOST:-$rgw_host}"
+
+# random argument determines if multipart is aborted or completed 50/50
+outcome=$((RANDOM % 2))
+if [ $outcome -eq 0 ] ;then
+    echo "== TESTING *ABORTING* MULTIPART UPLOAD WITH RE-UPLOADS =="
+else
+    echo "== TESTING *COMPLETING* MULTIPART UPLOAD WITH RE-UPLOADS =="
+fi
+
+# random argument determines if multipart is aborted or completed 50/50
+versioning=$((RANDOM % 2))
+if [ $versioning -eq 0 ] ;then
+    echo "== TESTING NON-VERSIONED BUCKET =="
+else
+    echo "== TESTING VERSIONED BUCKET =="
+fi
+
+# create a randomized bucket name
+bucket="reupload-bkt-$((RANDOM % 899999 + 100000))"
+
+
+# SET UP PYTHON VIRTUAL ENVIRONMENT
+
+# install boto3
+python3 -m venv $mydir
+source $mydir/bin/activate
+pip install pip --upgrade
+pip install boto3
+
+
+# CREATE RGW USER IF NECESSARY
+
+if radosgw-admin user info --access-key $RGW_ACCESS_KEY 2>/dev/null ;then
+    echo INFO: user already exists
+else
+    echo INFO: creating user
+    radosgw-admin user create --uid testid \
+                 --access-key $RGW_ACCESS_KEY \
+                 --secret $RGW_SECRET_KEY \
+                 --display-name 'M. Tester' \
+                 --email tester@ceph.com 2>/dev/null
+fi
+
+
+# RUN REUPLOAD TEST
+
+$mydir/bin/python3 ${mydir}/test_rgw_s3_mp_reupload.py $bucket $outcome $versioning
+
+
+# ANALYZE FOR ERRORS
+# (NOTE: for now we're choosing not to use the rgw-orphan-list tool)
+
+# force garbage collection to remove extra parts
+radosgw-admin gc process --include-all 2>/dev/null
+
+marker=$(radosgw-admin metadata get bucket:$bucket 2>/dev/null | grep bucket_id | sed 's/.*: "\(.*\)".*/\1/')
+
+# determine expected rados objects
+radosgw-admin bucket radoslist --bucket=$bucket 2>/dev/null | sort >$radoslist_out
+echo "radosgw-admin bucket radoslist:"
+cat $radoslist_out
+
+# determine found rados objects
+rados ls -p $data_pool 2>/dev/null | grep "^$marker" | sort >$rados_ls_out
+echo "rados ls:"
+cat $rados_ls_out
+
+# compare expected and found
+diff $radoslist_out $rados_ls_out >$diff_out
+if [ $(cat $diff_out | wc -l) -ne 0 ] ;then
+    error=1
+    echo "ERROR: Found differences between expected and actual rados objects for test bucket."
+    echo "    note: indicators: '>' found but not expected; '<' expected but not found."
+    cat $diff_out
+fi
+
+
+# CLEAN UP
+
+deactivate
+
+rm -f $orphan_list_out $radoslist_out $rados_ls_out $diff_out
+
+
+# PRODUCE FINAL RESULTS
+
+if [ -n "$error" ] ;then
+    echo "== FAILED =="
+    exit 1
+fi
+
+echo "== PASSED =="
+exit 0