Details
-
Bug
-
Resolution: Fixed
-
Major
-
7.6.0
-
OS - Debian 10
Couchbase server version - 7.6.0-1550
-
Untriaged
-
Linux x86_64
-
-
0
-
No
Description
Steps:
- Initialise a 2 node cluster running 7.6.0-1550 version with just the KV service. The 2 nodes are 172.23.106.242 and 172.23.107.110
- Configure magmaMinimum=30 in http://172.23.106.242:8091/settings/resourceManagement
{"bucket":{"collectionsPerQuota":{"enabled":true,"maximum":1},"dataSizePerNode":{"couchstoreMaximum":1.6,"enabled":true,"magmaMaximum":16},"residentRatio":{"couchstoreMinimum":10,"enabled":true,"magmaMinimum":30}},"coresPerBucket":{"enabled":true,"minimum":0.4},"diskUsage":{"enabled":true,"maximum":85}} |
- Create a Magma bucket named 'default' with ramQuota=512 per node. Create 3 scopes and 3 collections in each scope.
- Load data until the resident ratio guardrail (30%) is breached. Validated the same with some new inserts and received SDK error 'BUCKET_RESIDENT_RATIO_TOO_LOW'.
- Tried swap rebalance of 172.23.107.110 with a spare node 172.23.106.29 but it failed with the reason "rr_will_be_too_low:The following buckets are expected to breach the resident ratio minimum: default"
- The issue that I see here is, since swap rebalances do not ideally change the resident ratios of the bucket after the completion of rebalance, it should be allowed right?
Cb-collect info logs have been attached.