roachtest: ycsb/A/nodes=3 failed

[cockroach-teamcity]

SHA: https://github.com/cockroachdb/cockroach/commits/f28aa78febc98911122254a9d3c8212c0c940e08

Parameters:

To repro, try:

# Don't forget to check out a clean suitable branch and experiment with the
# stress invocation until the desired results present themselves. For example,
# using stress instead of stressrace and passing the '-p' stressflag which
# controls concurrency.
./scripts/gceworker.sh start && ./scripts/gceworker.sh mosh
cd ~/go/src/github.com/cockroachdb/cockroach && \
stdbuf -oL -eL \
make stressrace TESTS=ycsb/A/nodes=3 PKG=roachtest TESTTIMEOUT=5m STRESSFLAGS='-maxtime 20m -timeout 10m' 2>&1 | tee /tmp/stress.log

Failed test: https://teamcity.cockroachdb.com/viewLog.html?buildId=1158899&tab=buildLog

The test failed on release-2.1:
	cluster.go:1585,ycsb.go:89,ycsb.go:110,test.go:1211: ycsb/A/nodes=3 failed to meet throughput expectations: observed 2000, expected at least 498.3

[cockroach-teamcity]

SHA: https://github.com/cockroachdb/cockroach/commits/f28aa78febc98911122254a9d3c8212c0c940e08

Parameters:

To repro, try:

# Don't forget to check out a clean suitable branch and experiment with the
# stress invocation until the desired results present themselves. For example,
# using stress instead of stressrace and passing the '-p' stressflag which
# controls concurrency.
./scripts/gceworker.sh start && ./scripts/gceworker.sh mosh
cd ~/go/src/github.com/cockroachdb/cockroach && \
stdbuf -oL -eL \
make stressrace TESTS=ycsb/A/nodes=3 PKG=roachtest TESTTIMEOUT=5m STRESSFLAGS='-maxtime 20m -timeout 10m' 2>&1 | tee /tmp/stress.log

Failed test: https://teamcity.cockroachdb.com/viewLog.html?buildId=1158895&tab=buildLog

The test failed on release-2.1:
	cluster.go:1585,ycsb.go:89,ycsb.go:110,test.go:1211: ycsb/A/nodes=3 failed to meet throughput expectations: observed 2000, expected at least 371.3

[petermattis]

@ajwerner Looks like the failure message has expected and observed reversed. This failure is on 2.1. I believe the lower throughput is expected. Now that you've figured out the cause of the throughput drops on roachperf, perhaps we should just get rid of this minimal throughput check. The alternative is to provide different threshold values per release.