Skip to content

Commit 9ae16e8

Browse files
aleksbykovsoyacz
authored andcommitted
fix(perf_stress): Split stress multiplier by stress workload
Depend on stress workload, number of processes could give different max throughput. Split stress multiplier on 3 w,r,m. This will allow to configure different number of stress process on loader instance for write, read, mixed independently. (cherry picked from commit 2a5934e)
1 parent 06836cd commit 9ae16e8

File tree

2 files changed

+5
-7
lines changed

2 files changed

+5
-7
lines changed

docs/configuration_options.md

Lines changed: 1 addition & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -237,9 +237,7 @@
237237
| **<a href="#user-content-stress_multiplier_w" name="stress_multiplier_w">stress_multiplier_w</a>** | Number of cassandra-stress processes for write workload | 1 | SCT_STRESS_MULTIPLIER_W
238238
| **<a href="#user-content-stress_multiplier_r" name="stress_multiplier_r">stress_multiplier_r</a>** | Number of cassandra-stress processes for read workload | 1 | SCT_STRESS_MULTIPLIER_R
239239
| **<a href="#user-content-stress_multiplier_m" name="stress_multiplier_m">stress_multiplier_m</a>** | Number of cassandra-stress processes for mixed workload | 1 | SCT_STRESS_MULTIPLIER_M
240-
| **<a href="#user-content-run_fullscan" name="run_fullscan">run_fullscan</a>** | | N/A | SCT_RUN_FULLSCAN
241-
| **<a href="#user-content-run_full_partition_scan" name="run_full_partition_scan">run_full_partition_scan</a>** | Runs a background thread that issues reversed-queries on a table random partition by an interval | N/A | SCT_run_full_partition_scan
242-
| **<a href="#user-content-run_tombstone_gc_verification" name="run_tombstone_gc_verification">run_tombstone_gc_verification</a>** | Runs a background thread that verifies Tombstones GC on a table by an interval | N/A | SCT_RUN_TOMBSTONE_GC_VERIFICATION
240+
| **<a href="#user-content-run_fullscan" name="run_fullscan">run_fullscan</a>** | A list of dictionaries describing the parameters for the fullscan operations to be run. Each dictionary describes a separate thread to be spawned. Possible modes include: "table" for regular full table scans, "partition" for fullscans targeting partitions, "aggregate" for aggregate operations and "random" for a random selection of the former modes. | N/A | SCT_RUN_FULLSCAN
243241
| **<a href="#user-content-keyspace_num" name="keyspace_num">keyspace_num</a>** | | 1 | SCT_KEYSPACE_NUM
244242
| **<a href="#user-content-round_robin" name="round_robin">round_robin</a>** | | N/A | SCT_ROUND_ROBIN
245243
| **<a href="#user-content-batch_size" name="batch_size">batch_size</a>** | | 1 | SCT_BATCH_SIZE

test-cases/performance/perf-regression.100threads.30M-keys.yaml

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,9 @@
11
test_duration: 300
22

3-
stress_cmd_w: "cassandra-stress write no-warmup cl=QUORUM duration=60m -schema 'replication(strategy=NetworkTopologyStrategy,replication_factor=3)' -mode cql3 native -rate threads=100 -pop seq=1..30000000"
4-
prepare_write_cmd: "cassandra-stress write no-warmup cl=QUORUM n=30000000 -schema 'replication(strategy=NetworkTopologyStrategy,replication_factor=3)' -mode cql3 native -rate threads=100 -pop seq=1..30000000"
5-
stress_cmd_r: "cassandra-stress read no-warmup cl=QUORUM duration=50m -schema 'replication(strategy=NetworkTopologyStrategy,replication_factor=3)' -mode cql3 native -rate threads=100 -pop 'dist=gauss(1..30000000,15000000,1500000)' "
6-
stress_cmd_m: "cassandra-stress mixed no-warmup cl=QUORUM duration=50m -schema 'replication(strategy=NetworkTopologyStrategy,replication_factor=3)' -mode cql3 native -rate threads=100 -pop 'dist=gauss(1..30000000,15000000,1500000)' "
3+
stress_cmd_w: "cassandra-stress write no-warmup cl=QUORUM duration=60m -schema 'replication(factor=3)' -mode cql3 native -rate threads=100 -pop seq=1..30000000"
4+
prepare_write_cmd: "cassandra-stress write no-warmup cl=QUORUM n=30000000 -schema 'replication(factor=3)' -mode cql3 native -rate threads=100 -pop seq=1..30000000"
5+
stress_cmd_r: "cassandra-stress read no-warmup cl=QUORUM duration=50m -schema 'replication(factor=3)' -mode cql3 native -rate threads=100 -pop 'dist=gauss(1..30000000,15000000,1500000)' "
6+
stress_cmd_m: "cassandra-stress mixed no-warmup cl=QUORUM duration=50m -schema 'replication(factor=3)' -mode cql3 native -rate threads=100 -pop 'dist=gauss(1..30000000,15000000,1500000)' "
77
stress_multiplier_w: 2
88
stress_multiplier_r: 2
99
stress_multiplier_m: 2

0 commit comments

Comments
 (0)