[SERVER-31241] check compact behavior re: timestamps Created: 25/Sep/17  Updated: 06/Dec/22  Resolved: 29/Sep/17

Status: Closed
Project: Core Server
Component/s: Storage
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Major - P3
Reporter: Eric Milkie Assignee: Backlog - Storage Execution Team
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Assigned Teams:
Storage Execution
Participants:

 Description   

compact may or may not erase timestamp data on WiredTiger; we should check the behavior.



 Comments   
Comment by Eric Milkie [ 26/Sep/17 ]

No examples, we were just worried that something in the MongoDB code might affect timestamps. It's good to know nothing in the WiredTiger code will affect them.

Comment by Alexander Gorrod [ 26/Sep/17 ]

milkie The way compact works in WiredTiger is that it creates new checkpoints in a file using a different block allocation algorithm, and then reclaims unused space at the end of a database file. Since checkpoints can only be created "behind" the stable timestamp the way WiredTiger implements compaction should not overlap with timestamp semantics.

Have you seen an example where running compact alters the visibility of data?

Generated at Thu Feb 08 04:26:25 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.