bharatviswa504 edited a comment on pull request #1083:
URL: https://github.com/apache/hadoop-ozone/pull/1083#issuecomment-646220673


   @maobaolong 
   The issue with S3 patch is, it has just used fixed volume "/s3v", previously 
this volume is determined from accesskey, and it is a breaking change when 
upgrading from the previous version to the latest master/to be released 
version. As once after upgrade, users will not be able to access old buckets 
that they have created using the S3G. To handle that HDDS-3563 has done some 
code changes to make this configurable. So, that old buckets can be still 
accessible. 
   But if the previous cluster has n number of access key's we need a solution 
like s3 bind mount. 
   
   So, just exposing such a tool, and dropping a table from DB and upgrade the 
OM, will not work.  
   
   As currently, Ozone does has not have complete support for upgrades, once 
upgrade work is done, if any of these kinds of changes are performed, it is the 
developer's responsibility to handle these kinds of breaking changes. As said 
before, @avijayanhwx has discussed this in the upgrade design document in the 
section " Extending Layout Features for OM" which needs to be taken care of by 
LayoutFeature upgrade handlers. 
   
   So, in my view exposing such a tool from Ozone is not a good idea.
   
   
   >Admin move the rocksdb from one dir into another,
   >Admin backup&recover the rocksdb
   
   
   >Another truth, without this PR, i can still create/delete the rocksdb by 
rocksdb/ldb, with this PR, >i don't have to intall the rocksdb/ldb only.
   
   If an admin is doing these kinds of steps, it means he is risking data on 
the cluster.
   
   Finally, I want to say, any kind of incompatible changes should be taken 
care of by upgrade handlers to handle properly, and it should not require an 
admin to perform these kinds of operations. If it requires, that means our 
upgrade path is not correct.


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: ozone-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: ozone-issues-h...@hadoop.apache.org

Reply via email to