sql: disable buffered writes for implicit txns by default#164885
Open
yuzefovich wants to merge 1 commit intocockroachdb:masterfrom
Open
sql: disable buffered writes for implicit txns by default#164885yuzefovich wants to merge 1 commit intocockroachdb:masterfrom
yuzefovich wants to merge 1 commit intocockroachdb:masterfrom
Conversation
Contributor
|
Merging to
|
Member
|
This change is |
Member
🔴 Sysbench [SQL, 3node, oltp_read_write]
Reproducebenchdiff binaries: mkdir -p benchdiff/f87e0ff/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/f87e0ff60f1f86284babc4b2ca63bbfe5a2ee6d6/bin/pkg_sql_tests benchdiff/f87e0ff/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/f87e0ff/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/5bf7a3c/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/5bf7a3c99c92d1c5ed83c404df21301795fd6341/bin/pkg_sql_tests benchdiff/5bf7a3c/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/5bf7a3c/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/SQL/3node/oltp_read_write$ --old=5bf7a3c --new=f87e0ff --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_read_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/f87e0ff/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/f87e0ff60f1f86284babc4b2ca63bbfe5a2ee6d6/bin/pkg_sql_tests benchdiff/f87e0ff/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/f87e0ff/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/5bf7a3c/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/5bf7a3c99c92d1c5ed83c404df21301795fd6341/bin/pkg_sql_tests benchdiff/5bf7a3c/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/5bf7a3c/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_read_only$ --old=5bf7a3c --new=f87e0ff --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_write_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/f87e0ff/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/f87e0ff60f1f86284babc4b2ca63bbfe5a2ee6d6/bin/pkg_sql_tests benchdiff/f87e0ff/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/f87e0ff/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/5bf7a3c/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/5bf7a3c99c92d1c5ed83c404df21301795fd6341/bin/pkg_sql_tests benchdiff/5bf7a3c/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/5bf7a3c/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_write_only$ --old=5bf7a3c --new=f87e0ff --memprofile ./pkg/sql/testsArtifactsdownload: mkdir -p new
gcloud storage cp gs://cockroach-microbench-ci/artifacts/f87e0ff60f1f86284babc4b2ca63bbfe5a2ee6d6/22699013640-1/\* new/
mkdir -p old
gcloud storage cp gs://cockroach-microbench-ci/artifacts/5bf7a3c99c92d1c5ed83c404df21301795fd6341/22699013640-1/\* old/built with commit: f87e0ff60f1f86284babc4b2ca63bbfe5a2ee6d6 |
Member
Author
|
Hold off on reviewing - looks like I did something wrong. |
Member
Author
|
This is RFAL. The regression on sysbench was due to the special case I initially missed (that I mentioned in the commit message) where we might start an implicit txn for it to be upgraded soon after - this is now fixed by re-enabling buffered writes on the txn upgrade. |
Member
⚪ Sysbench [SQL, 3node, oltp_read_write]
Reproducebenchdiff binaries: mkdir -p benchdiff/92e4bfd/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/92e4bfdfe303a19e88974b50832a0404870aa3c6/bin/pkg_sql_tests benchdiff/92e4bfd/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/92e4bfd/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/SQL/3node/oltp_read_write$ --old=d423221 --new=92e4bfd --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_read_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/92e4bfd/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/92e4bfdfe303a19e88974b50832a0404870aa3c6/bin/pkg_sql_tests benchdiff/92e4bfd/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/92e4bfd/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_read_only$ --old=d423221 --new=92e4bfd --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_write_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/92e4bfd/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/92e4bfdfe303a19e88974b50832a0404870aa3c6/bin/pkg_sql_tests benchdiff/92e4bfd/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/92e4bfd/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_write_only$ --old=d423221 --new=92e4bfd --memprofile ./pkg/sql/testsArtifactsdownload: mkdir -p new
gcloud storage cp gs://cockroach-microbench-ci/artifacts/92e4bfdfe303a19e88974b50832a0404870aa3c6/23095406184-1/\* new/
mkdir -p old
gcloud storage cp gs://cockroach-microbench-ci/artifacts/d4232215214da4de11ec1ba71c88cd26a1f26573/23095406184-1/\* old/built with commit: 92e4bfdfe303a19e88974b50832a0404870aa3c6 |
Member
Author
|
Spoke too soon - there are some CI failures that I need to figure out. |
Member
⚪ Sysbench [SQL, 3node, oltp_read_write]
Reproducebenchdiff binaries: mkdir -p benchdiff/8764856/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/8764856ac514c540d1925e2269fa207327ffbbe4/bin/pkg_sql_tests benchdiff/8764856/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/8764856/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/SQL/3node/oltp_read_write$ --old=d423221 --new=8764856 --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_read_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/8764856/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/8764856ac514c540d1925e2269fa207327ffbbe4/bin/pkg_sql_tests benchdiff/8764856/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/8764856/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_read_only$ --old=d423221 --new=8764856 --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_write_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/8764856/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/8764856ac514c540d1925e2269fa207327ffbbe4/bin/pkg_sql_tests benchdiff/8764856/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/8764856/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_write_only$ --old=d423221 --new=8764856 --memprofile ./pkg/sql/testsArtifactsdownload: mkdir -p new
gcloud storage cp gs://cockroach-microbench-ci/artifacts/8764856ac514c540d1925e2269fa207327ffbbe4/23099007788-1/\* new/
mkdir -p old
gcloud storage cp gs://cockroach-microbench-ci/artifacts/d4232215214da4de11ec1ba71c88cd26a1f26573/23099007788-1/\* old/built with commit: 8764856ac514c540d1925e2269fa207327ffbbe4 |
|
Detected infrastructure failure (matched: self-hosted runner lost communication with the server, self-hosted runner lost communication with the server, self-hosted runner lost communication with the server). Automatically rerunning failed jobs. (run link) |
Member
⚪ Sysbench [SQL, 3node, oltp_read_write]
Reproducebenchdiff binaries: mkdir -p benchdiff/fbfc865/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/fbfc86599c113313add11a70dcd7589ee359d9a2/bin/pkg_sql_tests benchdiff/fbfc865/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/fbfc865/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/SQL/3node/oltp_read_write$ --old=d423221 --new=fbfc865 --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_read_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/fbfc865/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/fbfc86599c113313add11a70dcd7589ee359d9a2/bin/pkg_sql_tests benchdiff/fbfc865/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/fbfc865/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_read_only$ --old=d423221 --new=fbfc865 --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_write_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/fbfc865/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/fbfc86599c113313add11a70dcd7589ee359d9a2/bin/pkg_sql_tests benchdiff/fbfc865/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/fbfc865/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/d423221/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/d4232215214da4de11ec1ba71c88cd26a1f26573/bin/pkg_sql_tests benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/d423221/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_write_only$ --old=d423221 --new=fbfc865 --memprofile ./pkg/sql/testsArtifactsdownload: mkdir -p new
gcloud storage cp gs://cockroach-microbench-ci/artifacts/fbfc86599c113313add11a70dcd7589ee359d9a2/23117394708-1/\* new/
mkdir -p old
gcloud storage cp gs://cockroach-microbench-ci/artifacts/d4232215214da4de11ec1ba71c88cd26a1f26573/23117394708-1/\* old/built with commit: fbfc86599c113313add11a70dcd7589ee359d9a2 |
We expect that vast majority of implicit serializable txns don't benefit from the buffered writes feature while we've seen some regressions around blind writes, so this commit disables the buffered writes for all implicit txns. As an escape hatch, a new session variable `buffered_writes_implicit_txns_enabled` is introduced which can be set to `true` to get the previous behavior (to increase test coverage the default value for the session variable is metamorphically randomized in test builds). One notable case is worth calling out. In the PG wire protocol, entering the extended protocol starts an implicit transaction which is then upgraded to explicit if BEGIN is executed as a prepared statement. See 9fdb39b for more context. Release note: None
Member
⚪ Sysbench [SQL, 3node, oltp_read_write]
Reproducebenchdiff binaries: mkdir -p benchdiff/7b3a0e6/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/7b3a0e6e7173e1ee5ec0dd365937183d875e5d0b/bin/pkg_sql_tests benchdiff/7b3a0e6/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/7b3a0e6/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/74eff75/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/74eff753539d441da4247b99f3fd408a3708fe1e/bin/pkg_sql_tests benchdiff/74eff75/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/74eff75/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/SQL/3node/oltp_read_write$ --old=74eff75 --new=7b3a0e6 --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_read_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/7b3a0e6/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/7b3a0e6e7173e1ee5ec0dd365937183d875e5d0b/bin/pkg_sql_tests benchdiff/7b3a0e6/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/7b3a0e6/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/74eff75/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/74eff753539d441da4247b99f3fd408a3708fe1e/bin/pkg_sql_tests benchdiff/74eff75/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/74eff75/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_read_only$ --old=74eff75 --new=7b3a0e6 --memprofile ./pkg/sql/tests⚪ Sysbench [KV, 3node, oltp_write_only]
Reproducebenchdiff binaries: mkdir -p benchdiff/7b3a0e6/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/7b3a0e6e7173e1ee5ec0dd365937183d875e5d0b/bin/pkg_sql_tests benchdiff/7b3a0e6/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/7b3a0e6/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
mkdir -p benchdiff/74eff75/bin/1058449141
gcloud storage cp gs://cockroach-microbench-ci/builds/74eff753539d441da4247b99f3fd408a3708fe1e/bin/pkg_sql_tests benchdiff/74eff75/bin/1058449141/cockroachdb_cockroach_pkg_sql_tests
chmod +x benchdiff/74eff75/bin/1058449141/cockroachdb_cockroach_pkg_sql_testsbenchdiff command: # NB: for best (most stable) results, also add a suitable `--benchtime` that
# results in ~1s to ~5s of benchmark runs. For example, if ops average ~3ms, a
# benchtime of `1000x` is appropriate.
#
# Some benchmarks (in particular BenchmarkSysbench) output additional memory
# profiles covering only the execution (excluding the setup/teardown) - those
# should be preferred for analysis since they more closely correspond to what's
# reported as B/op and alloc/op.
benchdiff --run=^BenchmarkSysbench/KV/3node/oltp_write_only$ --old=74eff75 --new=7b3a0e6 --memprofile ./pkg/sql/testsArtifactsdownload: mkdir -p new
gcloud storage cp gs://cockroach-microbench-ci/artifacts/7b3a0e6e7173e1ee5ec0dd365937183d875e5d0b/23118556525-1/\* new/
mkdir -p old
gcloud storage cp gs://cockroach-microbench-ci/artifacts/74eff753539d441da4247b99f3fd408a3708fe1e/23118556525-1/\* old/built with commit: 7b3a0e6e7173e1ee5ec0dd365937183d875e5d0b |
|
Detected infrastructure failure (matched: self-hosted runner lost communication with the server). Automatically rerunning failed jobs. (run link) |
Member
Author
|
Ok, CI is green, so it's RFAL. Sorry for all the noise. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
We expect that vast majority of implicit serializable txns don't benefit from the buffered writes feature while we've seen some regressions around blind writes, so this commit disables the buffered writes for all implicit txns. As an escape hatch, a new session variable
buffered_writes_implicit_txns_enabledis introduced which can be set totrueto get the previous behavior (to increase test coverage the default value for the session variable is metamorphically randomized in test builds).One notable case is worth calling out. In the PG wire protocol, entering the extended protocol starts an implicit transaction which is then upgraded to explicit if BEGIN is executed as a prepared statement. See 9fdb39b for more context.
Fixes: #154553.
Release note: None