Replace Sled with other embedded databases #284
Labels
No Label
AdminAPI
Bug
Check AWS
CI
Correctness
Critical
Documentation
Ideas
Improvement
Low priority
Newcomer
Performance
S3 Compatibility
Testing
Usability
No Milestone
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: Deuxfleurs/garage#284
Loading…
Reference in New Issue
No description provided.
Delete Branch "%!s(<nil>)"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Sled is known to use lot of resources, including RAM. We thought we were able to cap RAM consumption by changing Sled's cache size. But it does not work in practise.
Investigations
It seems that cache size does not map to the effective RAM consumption:Also, it also seems that for some actions, it does not have any impact. For example, we are aware that for some people with 2GB of RAM, it is impossible to simply open a sled database.
Issues documenting Sled's memory problem: #986, #1093, #1061, #1304
Work in progress that may improve the situation: #1304
We are considering multiple options:
Note: in this specific case, this situation has been triggered because v0.6.0 had a bug and put 185M objects in the resync queue. Still, it should not prevent booting the server, the on-disk database being only 6GB large.
Sled memory usage might be out of controlto Replace Sled with another embedded databaseReplace Sled with another embedded databaseto Replace Sled with other embedded databases