Thread: timing sensitive failure of test_decoding RT that exists in PG9.6
timing sensitive failure of test_decoding RT that exists in PG9.6
From
"osumi.takamichi@fujitsu.com"
Date:
Hi
I by chance found a timing issue that exists in PG9.6
during my process to make a patch-set that should be back-patched to 9.6.
I don't judge this is unique in 9.6 or not from HEAD yet.
(Then, I don't know like this is solved by HEAD already but not back-patched but can it be ?)
Please tell me about this if you know something.
The logs I got :
* ./contrib/test_decoding/regression_output/regression.out
test ddl ... ok
test xact ... ok
test rewrite ... ok
test toast ... FAILED (test process exited with exit code 2)
test permissions ... FAILED (test process exited with exit code 2)
test decoding_in_xact ... FAILED (test process exited with exit code 2)
test decoding_into_rel ... FAILED (test process exited with exit code 2)
test binary ... FAILED (test process exited with exit code 2)
test prepared ... FAILED (test process exited with exit code 2)
test replorigin ... FAILED (test process exited with exit code 2)
test time ... FAILED (test process exited with exit code 2)
test messages ... FAILED (test process exited with exit code 2)
test spill ...
* ./contrib/test_decoding/regression_output/regression.diffs
*** /(where/I/put/PG)/contrib/test_decoding/expected/toast.out 2021-06-11 00:19:17.917565307 +0000
--- /(where/I/pug/PG)/contrib/test_decoding/./regression_output/results/toast.out 2021-06-11 00:37:45.642565307 +0000
***************
*** 348,364 ****
DROP TABLE toasted_several;
SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot',
NULL,NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
WHERE data NOT LIKE '%INSERT: %';
! regexp_replace
!
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
! BEGIN
! table public.toasted_several: UPDATE: old-key: id[integer]:1
toasted_key[text]:'98765432109876543210..7654321098765432109876543210987654321098765432109876543210'
toasted_col2[text]:unchanged-toast-datum
! table public.toasted_several: DELETE: id[integer]:1
toasted_key[text]:'98765432109876543210987654321..876543210987654321098765432109876543210987654321098765432109876543210987654321098765432109876543210'
! COMMIT
! (4 rows)
!
! SELECT pg_drop_replication_slot('regression_slot');
! pg_drop_replication_slot
! --------------------------
!
! (1 row)
!
--- 348,355 ----
DROP TABLE toasted_several;
SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot',
NULL,NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
WHERE data NOT LIKE '%INSERT: %';
! FATAL: could not open relation mapping file "global/pg_filenode.map": No such file or directory
! server closed the connection unexpectedly
! This probably means the server terminated abnormally
! before or while processing the request.
! connection to server was lost
....
I just checkouted the stable 9.6 branch
and configured with --enable-cassert --enable-debug --enable-tap-tests CFLAGS=-O0 --prefix=/where/I/put/binary.
Then, I ran build and make check-world in parallel. There is no core file.
Best Regards,
Takamichi Osumi