xpk/code-dumps
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

UPD: switched to pagination. resume with nextKey is dropped

Browse Source
This commit is contained in:
xpk 2024-07-16 13:14:33 +08:00
parent 77fc514056
commit 1afeac32f0
Signed by: xpk
GPG Key ID: CD4FF6793F09AB86
1 changed files with 53 additions and 42 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

95
py/s3-deep-archive-restore.py
View File

@ -1,9 +1,9 @@
#!/usr/bin/python3
import boto3
import pickle
import os
# import pickle
# import os
import concurrent.futures
from pprint import pprint
# from pprint import pprint
def restoreObject(bucket, key, versionId, myCount):
@ -30,49 +30,60 @@ bucketName = 's3-emr-hbase'
concurrency = 15
client = boto3.client('s3')
count = 0
try:
with open('nextKey.pickle', 'rb') as file:
nextKey = pickle.load(file)
response = client.list_object_versions(
Bucket=bucketName,
MaxKeys=10,
KeyMarker=nextKey
)
except IOError:
print("No position.pickle file. Start from beginning.")
response = client.list_object_versions(
Bucket=bucketName,
MaxKeys=10
)
nextKey = response.get("NextKeyMarker")
with concurrent.futures.ThreadPoolExecutor(max_workers=concurrency) as executor:
for i in response["Versions"]:
count += 1
# print(count, i.get("Key"), i.get("StorageClass"))
if i.get("StorageClass") == "DEEP_ARCHIVE":
executor.submit(restoreObject, bucketName, i.get("Key"), i.get("VersionId"), count)
print("NextKey", nextKey)
while nextKey is not None:
response = client.list_object_versions(
Bucket=bucketName,
MaxKeys=300,
KeyMarker=nextKey
)
nextKey = response.get("NextKeyMarker")
if nextKey is not None:
with open('nextKey.pickle', 'wb') as file:
pickle.dump(nextKey, file)
else:
os.remove('nextKey.pickle')
pagesize = 1000
paginator = client.get_paginator('list_object_versions')
result = paginator.paginate(Bucket=bucketName, PaginationConfig={'PageSize': pagesize})
for page in result:
with concurrent.futures.ThreadPoolExecutor(max_workers=concurrency) as executor:
for i in response["Versions"]:
for i in page["Versions"]:
count += 1
# print(count, i.get("Key"), i.get("StorageClass"))
if i.get("StorageClass") == "DEEP_ARCHIVE":
executor.submit(restoreObject, bucketName, i.get("Key"), i.get("VersionId"), count)
#
# try:
# with open('nextKey.pickle', 'rb') as file:
# nextKey = pickle.load(file)
# response = client.list_object_versions(
# Bucket=bucketName,
# MaxKeys=10,
# KeyMarker=nextKey
# )
# except IOError:
# print("No position.pickle file. Start from beginning.")
# response = client.list_object_versions(
# Bucket=bucketName,
# MaxKeys=10
# )
#
# nextKey = response.get("NextKeyMarker")
# with concurrent.futures.ThreadPoolExecutor(max_workers=concurrency) as executor:
# for i in response["Versions"]:
# count += 1
# # print(count, i.get("Key"), i.get("StorageClass"))
# if i.get("StorageClass") == "DEEP_ARCHIVE":
# executor.submit(restoreObject, bucketName, i.get("Key"), i.get("VersionId"), count)
#
# print("NextKey", nextKey)
# while nextKey is not None:
# response = client.list_object_versions(
# Bucket=bucketName,
# MaxKeys=300,
# KeyMarker=nextKey
# )
# nextKey = response.get("NextKeyMarker")
# if nextKey is not None:
# with open('nextKey.pickle', 'wb') as file:
# pickle.dump(nextKey, file)
# else:
# os.remove('nextKey.pickle')
# with concurrent.futures.ThreadPoolExecutor(max_workers=concurrency) as executor:
# for i in response["Versions"]:
# count += 1
# # print(count, i.get("Key"), i.get("StorageClass"))
# if i.get("StorageClass") == "DEEP_ARCHIVE":
# executor.submit(restoreObject, bucketName, i.get("Key"), i.get("VersionId"), count)
print('Total objects', count, sep=" ")