Hi, The probability of collision with MD5 is more compared to SHA256 based message digests. Even SHA128 is not strong enough and most dedupe engines use SHA256 based message digest.
Obviously if there is a collision, the incremental backup may not capture modified data which result in corrupt data backups. Murali Balcha > On Apr 11, 2017, at 3:52 AM, [email protected] wrote: > > hi,all > why we use sha256 instead md5 in chunk based backup-driver , in the > current file chunkeddriver.py, we have object_sha256['sha256s'] list to store > all the sha256 value of sha_block_size data , why we do not use md5 ,md5 is > faster compute than sha256,and we all use the md5 value in _backup_chunk > too, so we just need to compute one time and used in two diffrent place,it > save much time > > [email protected] > <mailto:[email protected]>_______________________________________________ > Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack > <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack> > Post to : [email protected] > <mailto:[email protected]> > Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack > <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack>
_______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : [email protected] Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
