Обсуждение: pgsql: In the pg_upgrade test suite, don't write to src/test/regress.
In the pg_upgrade test suite, don't write to src/test/regress. When this suite runs installcheck, redirect file creations from src/test/regress to src/bin/pg_upgrade/tmp_check/regress. This closes a race condition in "make -j check-world". If the pg_upgrade suite wrote to a given src/test/regress/results file in parallel with the regular src/test/regress invocation writing it, a test failed spuriously. Even without parallelism, in "make -k check-world", the suite finishing second overwrote the other's regression.diffs. This revealed test "largeobject" assuming @abs_builddir@ is getcwd(), so fix that, too. Buildfarm client REL_10, released forty-five days ago, supports saving regression.diffs from its new location. When an older client reports a pg_upgradeCheck failure, it will no longer include regression.diffs. Back-patch to 9.5, where pg_upgrade moved to src/bin. Reviewed by Andrew Dunstan. Discussion: https://postgr.es/m/20181224034411.GA3224776@rfd.leadboat.com Branch ------ master Details ------- https://git.postgresql.org/pg/commitdiff/bd1592e8570282b1650af6b8eede0016496daecd Modified Files -------------- src/bin/pg_upgrade/test.sh | 12 ++++++++++++ src/test/regress/input/largeobject.source | 6 +++--- src/test/regress/output/largeobject.source | 6 +++--- src/test/regress/output/largeobject_1.source | 6 +++--- src/tools/msvc/vcregress.pl | 18 +++++++++++++++++- 5 files changed, 38 insertions(+), 10 deletions(-)
Noah Misch <noah@leadboat.com> writes: > In the pg_upgrade test suite, don't write to src/test/regress. buildfarm is pretty unimpressed :-( regards, tom lane
Hi, On 2019-05-19 21:40:43 +0000, Noah Misch wrote: > In the pg_upgrade test suite, don't write to src/test/regress. > > When this suite runs installcheck, redirect file creations from > src/test/regress to src/bin/pg_upgrade/tmp_check/regress. This closes a > race condition in "make -j check-world". If the pg_upgrade suite wrote > to a given src/test/regress/results file in parallel with the regular > src/test/regress invocation writing it, a test failed spuriously. Even > without parallelism, in "make -k check-world", the suite finishing > second overwrote the other's regression.diffs. This revealed test > "largeobject" assuming @abs_builddir@ is getcwd(), so fix that, too. > > Buildfarm client REL_10, released forty-five days ago, supports saving > regression.diffs from its new location. When an older client reports a > pg_upgradeCheck failure, it will no longer include regression.diffs. > Back-patch to 9.5, where pg_upgrade moved to src/bin. > > Reviewed by Andrew Dunstan. The buildfarm decidedly is not a fan of this: https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=spurfowl&dt=2019-05-19%2021%3A43%3A32 Greetings, Andres Freund
Hi, On 2019-05-19 14:54:50 -0700, Andres Freund wrote: > On 2019-05-19 21:40:43 +0000, Noah Misch wrote: > > In the pg_upgrade test suite, don't write to src/test/regress. > > The buildfarm decidedly is not a fan of this: > https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=spurfowl&dt=2019-05-19%2021%3A43%3A32 And now also a different type of failure: https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=dory&dt=2019-05-19%2021%3A45%3A26 Last file mtime in snapshot: Sun May 19 21:36:44 2019 GMT =================================================== Can't locate object method "installcheck_internal" via package "serial" (perhaps you forgot to load "serial"?) at vcregress.plline 105. Greetings, Andres Freund
Andres Freund <andres@anarazel.de> writes: > The buildfarm decidedly is not a fan of this: > https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=spurfowl&dt=2019-05-19%2021%3A43%3A32 Also, dory is showing a different failure mode, in which it's not even getting to the pg_upgrade test: https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=dory&dt=2019-05-19%2021%3A45%3A26 The regular installcheck test is dying with Can't locate object method "installcheck_internal" via package "serial" (perhaps you forgot to load "serial"?) at vcregress.plline 105. Seems like some kind of syntax problem in the patch's change to vcregress.pl? I'm not enough of a Perl guy to diagnose it. regards, tom lane