Bug #40753
closed
Tasks #36451: mgr/dashboard: Scalability testing
mgr/dashboard: Perform scalability tests with large amounts of RGW buckets
Added by Lenz Grimmer almost 5 years ago.
Updated about 3 years ago.
Backport:
nautilus, octopus
Description
Evaluate the performance of the dashboard with increasing sizes of RGW buckets to identify potential bottlenecks and performance improvements.
- Copied to Tasks #40766: mgr/dashboard: Perform scalability tests with large amounts of RGW users added
- Tracker changed from Tasks to Bug
- Status changed from New to In Progress
- Assignee set to Ernesto Puerta
- Priority changed from Normal to Urgent
- Target version changed from v15.0.0 to v16.0.0
- Tags set to scalability
- Backport set to nautilus, octopus
- Regression set to No
- Severity set to 1 - critical
This has been reproduced in downstream (nautilus based) with more than 2k buckets and real network conditions, and basically renders RGW Bucket management unusable. Given that RGW bucket hasn't changed too much since Nautilus, this will likely be reproducible in upstream. I will try to.
Good news is that the back-end works and return the whole 2k list of buckets. Front-end reports net::ERR_INSUFFICIENT_RESOURCES
, so this might be due to the front-end launching many requests in parallel for each bucket.
The problem here is that the RGW Admin OPS API does not support filters and pagination. Because of that the Dashboard RGW controller need to retrieve ALL buckets, bypass them to the UI which then does the filtering.
I created an tracker issue looooong time ago that the Admin OPS API requires that feature: https://tracker.ceph.com/issues/22168
- Related to Feature #22168: The RGW Admin OPS is missing the ability to filter for e.g. buckets and users added
Volker Theile wrote:
The problem here is that the RGW Admin OPS API does not support filters and pagination. Because of that the Dashboard RGW controller need to retrieve ALL buckets, bypass them to the UI which then does the filtering.
I created an tracker issue looooong time ago that the Admin OPS API requires that feature: https://tracker.ceph.com/issues/22168
Thanks for the context, Volker!
- Related to Bug #47543: mgr/dashboard: fix performance issue when listing large amounts of buckets added
- Status changed from In Progress to Resolved
Also available in: Atom
PDF