-
Notifications
You must be signed in to change notification settings - Fork 4
perftest result history 930
sonyafenge edited this page Oct 7, 2022
·
13 revisions
- build branch: main - https://github.com/CentaurusInfra/global-resource-service
- Env for 1M:
export GRS_INSTANCE_PREFIX=grs-daily-0822-20x50 AUTORUN_E2E=true SIM_NUM=5 CLIENT_NUM=2 SERVER_NUM=1
export SERVER_ZONE=us-central1-a SIM_ZONE=us-central1-a,us-east1-b,us-west2-a,us-west4-a,us-west3-c CLIENT_ZONE=us-west3-b,us-east4-b
export SERVICE_EXTRA_ARGS="--enable_metrics=false"
export SIM_REGIONS="Beijing,Shanghai,Wulan,Guizhou,Reserved1" SIM_RP_NUM=8 NODES_PER_RP=25000 SCHEDULER_REQUEST_MACHINE=50000 SCHEDULER_REQUEST_LIMIT=51000 SCHEDULER_NUM=20
export SIM_DATA_PATTERN=Daily
- Env for 2M:
export GRS_INSTANCE_PREFIX=grs-daily-0822-40x50 AUTORUN_E2E=true SIM_NUM=5 CLIENT_NUM=2 SERVER_NUM=1
export SERVER_ZONE=us-central1-a SIM_ZONE=us-central1-a,us-east1-b,us-west2-a,us-west4-a,us-west3-c CLIENT_ZONE=us-west3-b,us-east4-b
export SERVICE_EXTRA_ARGS="--enable_metrics=false"
export SIM_REGIONS="Beijing,Shanghai,Wulan,Guizhou,Reserved1" SIM_RP_NUM=16 NODES_PER_RP=25000 SCHEDULER_REQUEST_MACHINE=50000 SCHEDULER_REQUEST_LIMIT=51000 SCHEDULER_NUM=40
export SIM_DATA_PATTERN=Daily
- Env for 5M:
export GRS_INSTANCE_PREFIX=grs-daily-1006-100x50 AUTORUN_E2E=true SIM_NUM=5 CLIENT_NUM=5 SERVER_NUM=1
export SERVER_ZONE=us-central1-a SIM_ZONE=us-central1-a,us-east1-b,us-west2-a,us-west4-a,us-west3-c CLIENT_ZONE=us-west3-b,us-west3-b,us-central1-b,us-east1-c,us-central1-a
export SERVICE_EXTRA_ARGS="--enable_metrics=false"
export SIM_REGIONS="Beijing,Shanghai,Wulan,Guizhou,Reserved1" SIM_RP_NUM=40 NODES_PER_RP=25000 SCHEDULER_REQUEST_MACHINE=50000 SCHEDULER_REQUEST_LIMIT=51000 SCHEDULER_NUM=100
export SIM_DATA_PATTERN=Daily,Daily,Daily,Daily,Daily SIM_WAIT_DOWN_TIME=5,5,5,5,5 SIM_DOWN_RP_NUM=0,0,0,0,0
export ENABLE_ADMIN_E2E=false
- Cmd:
./hack/test-setup.sh
- Log collection:
./hack/test-logcollect.sh
- Env for 5M
export GRS_INSTANCE_PREFIX=grs-down-1006-100x50 AUTORUN_E2E=true SIM_NUM=5 CLIENT_NUM=5 SERVER_NUM=1
export SERVER_ZONE=us-central1-a SIM_ZONE=us-central1-a,us-east1-b,us-west2-a,us-west4-a,us-west3-c CLIENT_ZONE=us-west3-b,us-west3-b,us-central1-b,us-east1-c,us-central1-a
export SERVICE_EXTRA_ARGS="--enable_metrics=false"
export SIM_REGIONS="Beijing,Shanghai,Wulan,Guizhou,Reserved1" SIM_RP_NUM=40 NODES_PER_RP=25000 SCHEDULER_REQUEST_MACHINE=50000 SCHEDULER_REQUEST_LIMIT=51000 SCHEDULER_NUM=100
export SIM_DATA_PATTERN=Outage,Outage,Outage,Outage,Outage SIM_WAIT_DOWN_TIME=7,7,7,7,7 SIM_DOWN_RP_NUM=40,40,40,40,40
export ENABLE_ADMIN_E2E=false
- Build and commit information: https://github.com/CentaurusInfra/global-resource-service/commit/80478ee939386b213e524c8d1c4889fe38083b36
80478ee (HEAD, upstream/main, sindica/main) Eligible sleep between simulator and scheduler, scheduler on same client machine (#151)
dbbdf33 Improvement for aggregator and simulator of Outage pattern (#150)
a0e1356 Issue 148 - code fix after merge. (#149)
31a7d40 Implement api to get node status (#147)
680f4da Reduce lock level and period for list (#146)
99d2195 Make maximum resource partition number per region 40. (#145)
...
- additional env var
--enable_metrics=false
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=40
SIM_RP_NUM=16
NODES_PER_RP=25000
- Test Result:
Test Case | Nodes Per RP | Scheduler # | Nodes per scheduler | Total Assigned Notes | Watch P50 (min/max-ms) | Watch P90 (min/max-ms) | Watch P99 (min/max-ms) | List (min/max-s) | Register (min/max-ms) | Metrics | Notes |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 25K | 40 (39) | 50K | 2M | 906/1050 | 1110/1151 | 1123/1163 | 1.39/1.57 | 59/1027 | N | N |
- Logs: Logs can be found at sonyadev4: /home/sonyali/grs/logs/1se5si2cl/082922-205134
- Build and commit information: https://github.com/CentaurusInfra/global-resource-service/commit/80478ee939386b213e524c8d1c4889fe38083b36
80478ee (HEAD, upstream/main, sindica/main) Eligible sleep between simulator and scheduler, scheduler on same client machine (#151)
dbbdf33 Improvement for aggregator and simulator of Outage pattern (#150)
a0e1356 Issue 148 - code fix after merge. (#149)
31a7d40 Implement api to get node status (#147)
680f4da Reduce lock level and period for list (#146)
99d2195 Make maximum resource partition number per region 40. (#145)
...
- additional env var
--enable_metrics=false
--enable_metrics=false
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
- Test Result:
Test Case | Nodes Per RP | Scheduler # | Nodes per scheduler | Total Assigned Notes | Watch P50 (min/max-ms) | Watch P90 (min/max-ms) | Watch P99 (min/max-ms) | List (min/max-s) | Register (min/max-ms) | Metrics | Notes |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 25K | 100 (99) | 50K | 5M | 723/998 | 997/1102 | 1020/1127 | 1.08/1.67 | 15/1300 | N | N |
- Logs: Logs can be found at sonyadev4: /home/sonyali/grs/logs/5000000/Outage/082922-224601
- Build and commit information: https://github.com/sonyafenge/global-resource-service/tree/perftest-930-0824
a1fdbf6 (HEAD -> pr-schedulerdelaystart, origin/pr-schedulerdelaystart) change scheduler delay time for 1M/2M
a2a5b07 (origin/pr-deslogdir, pr-deslogdir) change DES_LOG_DIR default value
80478ee (upstream/main, sidica/main) Eligible sleep between simulator and scheduler, scheduler on same client machine (#151)
dbbdf33 Improvement for aggregator and simulator of Outage pattern (#150)
a0e1356 Issue 148 - code fix after merge. (#149)
31a7d40 Implement api to get node status (#147)
680f4da Reduce lock level and period for list (#146)
99d2195 Make maximum resource partition number per region 40. (#145)
...
- additional env var
--enable_metrics=false
--enable_metrics=false
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=40
SIM_RP_NUM=16
NODES_PER_RP=25000
- Test Result:
Test Case | Nodes Per RP | Scheduler # | Nodes per scheduler | Total Assigned Notes | Watch P50 (min/max-ms) | Watch P90 (min/max-ms) | Watch P99 (min/max-ms) | List (min/max-s) | Register (min/max-ms) | Metrics | Notes |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 25K | 40 (39) | 50K | 2M | 854/974 | 1038/1119 | 1073/1139 | 1.36/1.59 | 59/83 | N | N |
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/2000000/Outage/083022-174806
09/06/2022 [grs][930][debug-test][Daily][5M][Single Region][Metrics]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/CentaurusInfra/global-resource-service/commit/4bf575c1fa61dfe4fdfd94494b77b248f0d4618e
4bf575c (HEAD, upstream/main) change DES_LOG_DIR default value (#153)
48fccfd Add GRS progress file
cd3c9c0 change scheduler delay time for 1M/2M (#154)
80478ee (sindica/main) Eligible sleep between simulator and scheduler, scheduler on same client machine (#151)
dbbdf33 Improvement for aggregator and simulator of Outage pattern (#150)
a0e1356 Issue 148 - code fix after merge. (#149)
31a7d40 Implement api to get node status (#147)
680f4da Reduce lock level and period for list (#146)
99d2195 Make maximum resource partition number per region 40. (#145)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
- Test Result:
Test Case | Nodes Per RP | Scheduler # | Nodes per scheduler | Total Assigned Notes | Watch P50 (min/max-ms) | Watch P90 (min/max-ms) | Watch P99 (min/max-ms) | List (min/max-s) | Register (min/max-ms) | Metrics | Notes |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 25K | 100 (99) | 50K | 5M | 7.11/10.22 | 11.51/13.95 | 13.25/22.50 | 1.11/1.42 | 15.11/55.01 | Y | N |
Metrics Item | perc50 | perc90 | perc99 | Total count |
---|---|---|---|---|
[Metrics][AGG_RECEIVED] | 7.63822ms | 11.929051ms | 13.302562ms | 5971 |
[Metrics][DIS_RECEIVED] | 7.646901ms | 11.941088ms | 13.311815ms | 5971 |
[Metrics][DIS_SENDING] | 7.849836ms | 12.170507ms | 13.536206ms | 5971 |
[Metrics][DIS_SENT] | 7.850541ms | 12.171248ms | 13.536883ms | 5971 |
[Metrics][SER_ENCODED] | 7.85924ms | 12.177806ms | 13.545743ms | 5971 |
[Metrics][SER_SENT] | 7.870677ms | 12.185894ms | 13.553078ms | 5971 |
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/Daily/090622-222451
09/06/2022 [grs][930][debug-test][Daily][5M][Multi Region][Metrics]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/CentaurusInfra/global-resource-service/commit/4bf575c1fa61dfe4fdfd94494b77b248f0d4618e
4bf575c (HEAD, upstream/main) change DES_LOG_DIR default value (#153)
48fccfd Add GRS progress file
cd3c9c0 change scheduler delay time for 1M/2M (#154)
80478ee (sindica/main) Eligible sleep between simulator and scheduler, scheduler on same client machine (#151)
dbbdf33 Improvement for aggregator and simulator of Outage pattern (#150)
a0e1356 Issue 148 - code fix after merge. (#149)
31a7d40 Implement api to get node status (#147)
680f4da Reduce lock level and period for list (#146)
99d2195 Make maximum resource partition number per region 40. (#145)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
- Test Result:
Test Case | Nodes Per RP | Scheduler # | Nodes per scheduler | Total Assigned Notes | Watch P50 (min/max-ms) | Watch P90 (min/max-ms) | Watch P99 (min/max-ms) | List (min/max-s) | Register (min/max-ms) | Metrics | Notes |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 25K | 100 (99) | 50K | 5M | 55.91/92.67 | 91.09/127.11 | 110.13/136.42 | 1.13/1.79 | 14.93/134.89 | Y | N |
Metrics Item | perc50 | perc90 | perc99 | Total count |
---|---|---|---|---|
[Metrics][AGG_RECEIVED] | 65.372751ms | 100.499979ms | 117.034877ms | 5947 |
[Metrics][DIS_RECEIVED] | 65.384097ms | 100.509367ms | 117.057475ms | 5947 |
[Metrics][DIS_SENDING] | 65.75874ms | 100.912655ms | 117.389416ms | 5947 |
[Metrics][DIS_SENT] | 65.759878ms | 100.91348ms | 117.390268ms | 5947 |
[Metrics][SER_ENCODED] | 65.766027ms | 100.92214ms | 117.398885ms | 5947 |
[Metrics][SER_SENT] | 65.775296ms | 100.931097ms | 117.405457ms | 5947 |
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/Daily/090622-235623
09/13/2022 [grs][930][debug-test][Daily] [5M][PR163][SingleRegion]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/CentaurusInfra/global-resource-service/pull/163
ffe6e75 (HEAD, yb01/sim-lw) fix flush time bug to avoid long delays for smaller changes
aee062d (yb01/sim-lw2) initial enabling of concurrent processing in lw
d6bc224 (upstream/simulator-list-watch) list-watch approach from simulator to aggregator (#160)
fa1c6d3 Revert "Temporary enable all watch latency for perf details."
b3b9f0d Temporary enable all watch latency for perf details.
187dd8a Aggregator list watch from simulator (#159)
4bf575c (yb01/perf, yb01/main, carl/main) change DES_LOG_DIR default value (#153)
48fccfd Add GRS progress file
cd3c9c0 change scheduler delay time for 1M/2M (#154)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/Daily/091322-215232
09/13/2022 [grs][930][debug-test][Daily] [5M][PR163][crossRegion]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/CentaurusInfra/global-resource-service/pull/163
ffe6e75 (HEAD, yb01/sim-lw) fix flush time bug to avoid long delays for smaller changes
aee062d (yb01/sim-lw2) initial enabling of concurrent processing in lw
d6bc224 (upstream/simulator-list-watch) list-watch approach from simulator to aggregator (#160)
fa1c6d3 Revert "Temporary enable all watch latency for perf details."
b3b9f0d Temporary enable all watch latency for perf details.
187dd8a Aggregator list watch from simulator (#159)
4bf575c (yb01/perf, yb01/main, carl/main) change DES_LOG_DIR default value (#153)
48fccfd Add GRS progress file
cd3c9c0 change scheduler delay time for 1M/2M (#154)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/Daily/091322-234340
- Build and commit information: https://github.com/CentaurusInfra/global-resource-service/commit/c1ceb297b7e53e790ccbbaff4b1d00d3d9dd13ac
c1ceb29 (HEAD, upstream/simulator-list-watch) disable scheduler watch delay log for regular perf runs (#166)
51e06bf initial enabling of concurrent processing in lw (#163)
d6bc224 list-watch approach from simulator to aggregator (#160)
fa1c6d3 Revert "Temporary enable all watch latency for perf details."
b3b9f0d Temporary enable all watch latency for perf details.
187dd8a Aggregator list watch from simulator (#159)
4bf575c (yb01/perf, yb01/main, carl/main) change DES_LOG_DIR default value (#153)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/Outage/091422-185047
- Build and commit information: https://github.com/sonyafenge/global-resource-service/tree/perftest-930-0914
761fa71 (HEAD, origin/perftest-930-0914) change to wait 90s to get all vms ready
3d2fdf6 Revert "remove instance group from test setup (#162)"
86e3c2e (origin/pr-listwatchmerge) disable scheduler watch delay log for regular perf runs (#166)
4e827e9 initial enabling of concurrent processing in lw (#163)
373d97e list-watch approach from simulator to aggregator (#160)
0ee44f6 Revert "Temporary enable all watch latency for perf details."
ecd7e23 Temporary enable all watch latency for perf details.
9fc1e0d Aggregator list watch from simulator (#159)
55749fa (upstream/main) Add function BatchLogicalNodesInquiry for redis store (#155)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/Daily/091422-210046
09/14/2022 [grs][930][Nodequery with batch][Daily][metrics] [5M][LS][CrossRegion]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/sonyafenge/global-resource-service/tree/perftest-930-0914
761fa71 (HEAD, origin/perftest-930-0914) change to wait 90s to get all vms ready
3d2fdf6 Revert "remove instance group from test setup (#162)"
86e3c2e (origin/pr-listwatchmerge) disable scheduler watch delay log for regular perf runs (#166)
4e827e9 initial enabling of concurrent processing in lw (#163)
373d97e list-watch approach from simulator to aggregator (#160)
0ee44f6 Revert "Temporary enable all watch latency for perf details."
ecd7e23 Temporary enable all watch latency for perf details.
9fc1e0d Aggregator list watch from simulator (#159)
55749fa (upstream/main) Add function BatchLogicalNodesInquiry for redis store (#155)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/Daily/091422-222008
- Build and commit information: https://github.com/sonyafenge/global-resource-service/tree/perftest-930-0916
304801e (HEAD -> perftest-930-0916, origin/perftest-930-0916) Should not flushall redis DB when service API starts
60cd037 Change Redis default port to new port
8c177bc (origin/pr-simdiffdatapattern) fix parameter change for wait_time_for_data_change_pattern
306d7f6 support different data pattern for each simulator
ef4d62e (upstream/simulator-list-watch) change delay to 90s to get all vms ready in test-setup (#177)
33f4e2d changes to test throughput (#175)
8b3806a Revert "remove instance group from test setup (#162)" (#169) (#173)
1a244ca Merge main into simulator-list-watch branch (#170)
d28fafe Add sleep time for Daily data pattern (#165)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
SIM_DATA_PATTERN=Daily,Outage,Daily,Daily,Daily
SIM_WAIT_DOWN_TIME=0,6,0,0,0
SIM_DOWN_RP_NUM=0,1,0,0,0
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/091622-202259
09/16/2022 [grs][930][debug-test][MIX] [5M][Throughput][10RPDown]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/sonyafenge/global-resource-service/tree/perftest-930-0916
304801e (HEAD -> perftest-930-0916, origin/perftest-930-0916) Should not flushall redis DB when service API starts
60cd037 Change Redis default port to new port
8c177bc (origin/pr-simdiffdatapattern) fix parameter change for wait_time_for_data_change_pattern
306d7f6 support different data pattern for each simulator
ef4d62e (upstream/simulator-list-watch) change delay to 90s to get all vms ready in test-setup (#177)
33f4e2d changes to test throughput (#175)
8b3806a Revert "remove instance group from test setup (#162)" (#169) (#173)
1a244ca Merge main into simulator-list-watch branch (#170)
d28fafe Add sleep time for Daily data pattern (#165)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
SIM_DATA_PATTERN=Daily,Outage,Daily,Daily,Daily
SIM_WAIT_DOWN_TIME=4,7,4,4,4
SIM_DOWN_RP_NUM=0,10,0,0,0
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/091622-222458
09/16/2022 [grs][930][debug-test][MIX] [5M][Throughput][40RPDown]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/sonyafenge/global-resource-service/tree/perftest-930-0916
304801e (HEAD -> perftest-930-0916, origin/perftest-930-0916) Should not flushall redis DB when service API starts
60cd037 Change Redis default port to new port
8c177bc (origin/pr-simdiffdatapattern) fix parameter change for wait_time_for_data_change_pattern
306d7f6 support different data pattern for each simulator
ef4d62e (upstream/simulator-list-watch) change delay to 90s to get all vms ready in test-setup (#177)
33f4e2d changes to test throughput (#175)
8b3806a Revert "remove instance group from test setup (#162)" (#169) (#173)
1a244ca Merge main into simulator-list-watch branch (#170)
d28fafe Add sleep time for Daily data pattern (#165)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
SIM_DATA_PATTERN=Daily,Outage,Daily,Daily,Daily
SIM_WAIT_DOWN_TIME=4,7,4,4,4
SIM_DOWN_RP_NUM=0,40,0,0,0
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/091622-234845
09/19/2022 [grs][930][debug-test][MIX] [5M][Throughput][200RPDown]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/sonyafenge/global-resource-service/tree/perftest-930-0919
7cdd28d (HEAD -> perftest-930-0916) fix grep error for scv report
e935fd4 add throughput data to scv report
f856661 (origin/perftest-930-0919) Allow watch multiple regions in scheduler
304801e (origin/perftest-930-0916) Should not flushall redis DB when service API starts
60cd037 Change Redis default port to new port
8c177bc fix parameter change for wait_time_for_data_change_pattern
306d7f6 support different data pattern for each simulator
ef4d62e (upstream/simulator-list-watch) change delay to 90s to get all vms ready in test-setup (#177)
33f4e2d changes to test throughput (#175)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
SIM_DATA_PATTERN=Outage,Outage,Outage,Outage,Outage
SIM_WAIT_DOWN_TIME=7,7,7,7,7
SIM_DOWN_RP_NUM=40,40,40,40,40
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/091922-184339
10/06/2022 [grs][930][release][Mix-Dailyonly] [5M][watchdelaylog]scheduler100x50K, simulator5x40x25k
- Build and commit information: https://github.com/sonyafenge/global-resource-service/tree/perftest-930-1006-daily
f7e7dcf (HEAD -> perftest-930-1006-daily, origin/perftest-930-1006-daily) Add all watch delay into log
bc1c40f (upstream/main) Simulator list watch merge (#196)
85102f7 move nodequery from test/e2e to resource-management/test/e2e (#188)
ef3e874 (sindica/main) Change Redis default port 6379 to new port flexiably (#171)
7e3abc0 Should not flushall redis DB when service API starts (#167)
94f5cba Update UT performance data for main/PR 169 (#185)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
SIM_DATA_PATTERN=Daily,Daily,Daily,Daily,Daily
SIM_WAIT_DOWN_TIME=5,5,5,5,5
SIM_DOWN_RP_NUM=0,0,0,0,0
- Results:
Perc 50: 11.83ms/47.72ms
Perc 90: 51.54ms/88.51ms
Perc 99: 72.36ms/90.05ms
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/100622-235002
- Build and commit information: https://github.com/CentaurusInfra/global-resource-service
bc1c40f (HEAD, upstream/main) Simulator list watch merge (#196)
85102f7 move nodequery from test/e2e to resource-management/test/e2e (#188)
ef3e874 (sindica/main) Change Redis default port 6379 to new port flexiably (#171)
7e3abc0 Should not flushall redis DB when service API starts (#167)
94f5cba Update UT performance data for main/PR 169 (#185)
...
- additional env var
SCHEDULER_REQUEST_MACHINE=50000
SCHEDULER_NUM=100
SIM_RP_NUM=40
NODES_PER_RP=25000
SIM_DATA_PATTERN=Outage,Outage,Outage,Outage,Outage
SIM_WAIT_DOWN_TIME=7,7,7,7,7
SIM_DOWN_RP_NUM=40,40,40,40,40
- Results:
Perc 50: 21.27s/25.25s
Perc 90: 43.86s/45.19s
Perc 99: 47.59s/47.82s
Throughput | ||
---|---|---|
File Name | Items | Time |
grs-down-1006-100x50-client-us-central1-b-2.log.4 | THP_START_MIN | 10/7/22 1:11:21.387 |
grs-down-1006-100x50-client-us-east1-c-3.log.9 | THP_END_MAX | 10/7/22 1:12:13.340 |
- Logs: Logs can be found at sonyadev4:/home/sonyali/grs/logs/5000000/100722-015924