lpavanvenkat commented on PR #11685: URL: https://github.com/apache/ozone/pull/11685#issuecomment-6595587820
> Thanks @lpavanvenkat for the clarification. That explanation makes sense. It looks different than the PR description though: > > > for one MPU with 10 MiB of committed parts: > > Endpoint Before After > /keys/open row 0 10 MiB > /keys/open/summary 0 10 MiB > /keys/open/mpu/summary 10 MiB 10 MiB (unchanged) > This seems to indicate that the 10mb of committed but not completed parts were not counted towards open key size before this PR, but that they are after the PR. Your latest comment says that the 10mb of completed parts would not count towards the open key size. > > > For example, start 2 multipart uploads and commit 2 parts of 5 MiB to each, without completing either upload. Recon then reports: > > > /keys/open/summary: 2 open keys, 0 bytes Thanks @errose28, you're right, they don't match. The PR description describes the current code, which adds committed part sizes to /keys/open and /keys/open/summary. The proposal replaces that: committed parts are counted only in /keys/open/mpu/summary, and the size-0 multipart open key is excluded from /keys/open and /keys/open/summary. For one MPU with 10 MiB of committed parts and no part being written, the proposal would give: | Endpoint | Before | After | | -- | -- | -- | | /keys/open/summary | 1 open key, 0 bytes | 0 open keys, 0 bytes | | /keys/open | 1 row, 0 bytes | no row | | /keys/open/mpu/summary | 1 upload, 10 MiB | 1 upload, 10 MiB (unchanged) | If this works for you, I'll rework the PR to this approach and update the description. -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
