[SERVER-40799] Use collection UUID in WiredTiger collection URIs instead of random number Created: 24/Apr/19  Updated: 06/Dec/22  Resolved: 27/Apr/20

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

Type: Improvement Priority: Major - P3
Reporter: Geert Bosch Assignee: Backlog - Storage Execution Team
Resolution: Won't Do Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Assigned Teams:
Storage Execution
Participants:

 Description   

Currently we use names like collection-0--7004555647483938944.wt, where the part preceding the . is a signed 64-bit random number. These names are hard to correlate with user collections. Use the UUID instead to make this easier.

This change will only affect newly created files.



 Comments   
Comment by Geert Bosch [ 14/Jun/19 ]

Note that for the indexes we still need the random numbers. If we're implementing this ticket, we may need to ensure that recreating an ident with the same UUID/options will work (and just truncate the table), so rollback will work correctly.

Comment by Daniel Gottlieb (Inactive) [ 24/Apr/19 ]

FWIW, I have used both the auto-incrementing piece as well as the random number generation ("process id") piece in diagnostics before.

Generated at Thu Feb 08 04:56:00 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.