Sachin Jayant Joshi

I'm not sure how this could be possible. There is no code path that adjusts metadata length of last chunk except inside `claimOwnership` which is now turned off by default....

Also I think to just avoid any future confusion let's always turn off optimization on system segments and any attribute segments irrespective of whether lazy commit set or not.

With no optimization, for each LTS write we'll update and commit 2 or 3 metadata records - record for segment , record for last chunk and record for new chunk...

@AJadhav29 @Amit-Singh40 Please review this change. Thanks

https://github.com/pravega/pravega/pull/6909 For porting changes to 0.12 branch

I tried multiple times making one big change that removes all references to RollingStorage . It becomes a very big change. With some tests still failing. Therefore I'm taking incremental...

@Amit-Singh40 @AJadhav29 Please review.

System tests passed with ``` storage.self.check.integrity.data=true storage.self.check.integrity.metadata=true ```

@tkaitchuck I addressed the review comments. Please take another look. Thanks

@abhinb This implementation is too specific to FileSystem. What happens with ECS, S3 or other chunk storages? Also if you could move some of the heavy functionality to https://github.com/pravega/pravega/blob/master/segmentstore/storage/src/main/java/io/pravega/segmentstore/storage/chunklayer/UtilsWrapper.java That...