Обсуждение: pgsql: logical decoding: Tell reorderbuffer about all xids.
logical decoding: Tell reorderbuffer about all xids. Logical decoding's reorderbuffer keeps transactions in an LSN ordered list for efficiency. To make that's efficiently possible upper-level xids are forced to be logged before nested subtransaction xids. That only works though if these records are all looked at: Unfortunately we didn't do so for e.g. row level locks, which are otherwise uninteresting for logical decoding. This could lead to errors like: "ERROR: subxact logged without previous toplevel record". It's not sufficient to just look at row locking records, the xid could appear first due to a lot of other types of records (which will trigger the transaction to be marked logged with MarkCurrentTransactionIdLoggedIfAny). So invent infrastructure to tell reorderbuffer about xids seen, when they'd otherwise not pass through reorderbuffer.c. Reported-By: Jarred Ward Bug: #13844 Discussion: 20160105033249.1087.66040@wrigleys.postgresql.org Backpatch: 9.4, where logical decoding was added Branch ------ master Details ------- http://git.postgresql.org/pg/commitdiff/d9e903f3cbbd00c7ba7d4974e6852c3d2cbf4447 Modified Files -------------- contrib/test_decoding/Makefile | 4 +-- contrib/test_decoding/expected/xact.out | 42 +++++++++++++++++++++++++ contrib/test_decoding/sql/xact.sql | 22 +++++++++++++ src/backend/replication/logical/decode.c | 28 ++++++++++++++++- src/backend/replication/logical/reorderbuffer.c | 21 ++++++++----- src/backend/replication/logical/snapbuild.c | 6 +--- src/include/replication/reorderbuffer.h | 2 +- 7 files changed, 108 insertions(+), 17 deletions(-)
Andres Freund <andres@anarazel.de> writes: > logical decoding: Tell reorderbuffer about all xids. One or another of these logical-decoding fixes is making all the ppc64 buildfarm members unhappy. regards, tom lane
On 2016-03-05 22:49:18 -0500, Tom Lane wrote: > Andres Freund <andres@anarazel.de> writes: > > logical decoding: Tell reorderbuffer about all xids. > > One or another of these logical-decoding fixes is making all the ppc64 > buildfarm members unhappy. Just saw, looking into it. Might take me till tomorrow morning to finish. It's very likely the last of the three commits, but we'll see. Andres
On 2016-03-05 22:49:18 -0500, Tom Lane wrote: > Andres Freund <andres@anarazel.de> writes: > > logical decoding: Tell reorderbuffer about all xids. > > One or another of these logical-decoding fixes is making all the ppc64 > buildfarm members unhappy. http://pgbuildfarm.org/cgi-bin/show_log.pl?nm=termite&dt=2016-03-07%2000%3A37%3A22 Looks like my attempt at fixing this was successful. While debugging this I found some minor further improvements, which I'll fix together with the issue coverity spotted. Andres
Andres Freund <andres@anarazel.de> writes: > On 2016-03-05 22:49:18 -0500, Tom Lane wrote: >> One or another of these logical-decoding fixes is making all the ppc64 >> buildfarm members unhappy. > http://pgbuildfarm.org/cgi-bin/show_log.pl?nm=termite&dt=2016-03-07%2000%3A37%3A22 > Looks like my attempt at fixing this was successful. Yeah, gaur is happy too. No ARM reports yet though :-( regards, tom lane