Details
-
Bug
-
Resolution: Fixed
-
Major
-
6.5.0
-
Untriaged
-
Yes
Description
Observed degradation in rebalance time for graceful failover scenario
Test : Graceful failover (ms), 4 -> 3, 1 bucket x 100M x 2KB, 10K ops/sec
Test | Build 6.5.0-3789 | Build 6.5.0-3790 |
Rebalance Time(ms) | 3457 | 6166 |
http://showfast.sc.couchbase.com/#/timeline/Linux/reb/failover/all
Looks like degradation introduced due to below change in kv_engine . Please take a look .
changed projects :
|
|
kv_engine changed from 70a4d0af1d1c55d1494a1b777ed7935c34c36dc6 to master
|
[+] bc29fba MB-35096: Check received seqnoAck <= highest sent item
|
[+] 9bfe8e2 MB-35053: Extend allowedDuplicatePrepareSeqnos to highSeqno
|
Attachments
Issue Links
- is duplicated by
-
MB-35316 [System Test] Memcached crashes with status 137
- Closed
For Gerrit Dashboard: MB-35252 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
112596,6 | MB-35252: Simplify allowed duplicate prepare logic | master | kv_engine | Status: MERGED | +2 | +1 |