[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ceph-users
Subject:    [ceph-users] Write Order during Concurrent S3 PUT on RGW
From:       Scheurer_François <francois.scheurer () everyware ! ch>
Date:       2021-09-29 14:54:21
Message-ID: 45804eb26705454298da68fa79a9609a () everyware ! ch
[Download RAW message or body]

This is an S/MIME signed message

--===============3807671546259668534==
Content-Language: en-US
Content-Type: multipart/signed; protocol="application/pkcs7-signature"; \
micalg="sha-256"; boundary="----21514C5B5465F9A4F9B99B7D8DE2DF0A"

This is an S/MIME signed message



Dear All,



RGW provides atomic PUT in order to guarantee write consistency.

cf: https://ceph.io/en/news/blog/2011/atomicity-of-restful-radosgw-operatio=
ns/

But my understanding is that the are no guarantee regarding the PUT order s=
equence.
So basically, if doing a storage class migration:
aws s3 cp s3://<bucket-name>/ s3://<bucket-name>/ --recursive --storage-cla=
ss <storage_class>
Then if another client is writing concurrently in this bucket, a PUT collis=
ion could happen.
Is it correct that in this case the slowest PUT would win?
In fact, if the other client starts its PUT after my PUT starts and before =
my PUT ends,
then my PUT would win, overwriting new data with old data...

Any ideas to prevent that?
Maybe with "radosgw-admin bucket rewrite" ?
Maybe by "locking" the object temporary with an ACL or something else?

This potential issue could also affect LifeCycle Policies that use StorageC=
lass Transition...

Thank you for your help.


Cheers
Francois


--


EveryWare AG
Fran=E7ois Scheurer
Senior Systems Engineer
Zurlindenstrasse 52a
CH-8003 Z=FCrich

tel: +41 44 466 60 00
fax: +41 44 466 60 10
mail: francois.scheurer@everyware.ch
web: http://www.everyware.ch


_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-leave@ceph.io

--===============3807671546259668534==--


[prev in list] [next in list] [prev in thread] [next in thread]