Details
-
Improvement
-
Resolution: Done
-
Major
-
6.5.0
Description
At the moment we have improved backed up to a throughput of around 400MB/s but restore is still under 300MB/s.
From running restore to blachole we can see that some of the slowness does not come from the archive source as I originally thought but instead some of it is attributed to the couchbase sink as the through put for a restore to the cluster with 1 thread is around 30 MB/s but to blackhole is around 90 MB/s.
Attachments
Issue Links
- relates to
-
MB-34396 CBM restore to blackhole
- Closed
For Gerrit Dashboard: MB-34397 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
119101,13 | MB-34397 Refactored the restore pipeline | master | backup | Status: ABANDONED | +1 | +1 |
121340,11 | MB-34397 Refactored the restore worker pool | master | backup | Status: MERGED | +2 | +1 |