Couchbase Server
Performance Status
05/21/2014
KV
MB-11143:
Avg. BgFetcher wait time is still 5-6 times higher in 3.0 (on slow HDD)
MB-11083:
Very high CPU utilization (1700%) and growing disk queue when compaction is running
MB-11005:
Erlang memory in KV test case with 10 buckets is too high (up to 30GB, leading to OOM situation)
MB-10771: Delta recovery is slower than full recovery. Meanwhile several performance issues encountered.
KV
MB-10679: It takes almost 10 minutes to stop Couchbase service in setup with 10 empty buckets
MB-10370:
ep-engine deadlock in write-heavy DGM cases
MB-9676:
Dramatically increasing latency of SET operations during rebalance tests
MB-9620: (or just kill mccouch) multi-tenancy: beam.smp memory usage optimization for non-views cases
MB-9461:
In heavy-DGM (<5%) scenarios with views and high cache miss rate rebalance/client operations can fail due to timed out requests to memcached/ep-engine
XDCR
-
MB-11058:
Failover during data load with enabled XDCR
-
MB-10943:
Erlang memory usage goes up to 60GB in XDCR setups [was: Source node auto failed over during initial data load with XDCR]
-
MB-10437: XDCR replication rate drops almost to zero in presence of light write workload on src side
-
MB-7907: Issues when scaling XDCR on single node
Views
-
MB-10956:
Rebalance-in with views takes 10 hours (used to be 70 minutes)
-
MB-10273:
View compaction doesn't catch up in basic non-DGM tests with view queries
-
MB-9822:
One of nodes is too slow during indexing
Windows
MB-10191: CouchDB crashed due to 'Cannot allocate 467078560 bytes of memory (of type "heap").'
MB-9825: Rebalance exited with reason bad_replicas
n1ql
-
MB-11141: SELECT COUNT(*) ... WHERE should avoid memcached operations
-
MB-11140: SELECT DISTINCT scans all docs
-
MB-11048: Range queries result in thousands of GET operations/sec
BACKLOG
-
[ongoing] Query performance testing
- More "observe" tests
- Mobile dashboard
- SmartOS + DTrace
CBIT ZONE
-
CBIT-1182:
client machines for regression tests
-
CBIT-1158:
RAID 10 drives for performance testing