Hi! Thanks for new patches. Yet another crash reproduced on master with v63 patches:
CREATE TABLESPACE ts LOCATION '/tmp/ts'; CREATE GLOBAL TEMP TABLE tbl (num1 bigint); INSERT INTO tbl (num1) values (1); CREATE INDEX tbl_idx ON tbl (num1); REINDEX (TABLESPACE ts) TABLE tbl; Got error: CREATE TABLESPACE CREATE TABLE INSERT 0 1 CREATE INDEX WARNING: AbortTransaction while in COMMIT state ERROR: gtt relfilenode 16388 not found in rel 16388 PANIC: cannot abort transaction 726, it was already committed server closed the connection unexpectedly This probably means the server terminated abnormally before or while processing the request. connection to server was lost in log: 2021-12-21 12:54:08.273 +07 [208725] ERROR: gtt relfilenode 16388 not found in rel 16388 2021-12-21 12:54:08.273 +07 [208725] STATEMENT: REINDEX (TABLESPACE ts) TABLE tbl; 2021-12-21 12:54:08.273 +07 [208725] WARNING: AbortTransaction while in COMMIT state 2021-12-21 12:54:08.273 +07 [208725] PANIC: cannot abort transaction 726, it was already committed 2021-12-21 12:54:08.775 +07 [208716] LOG: server process (PID 208725) was terminated by signal 6: Аварийный останов 2021-12-21 12:54:08.775 +07 [208716] DETAIL: Failed process was running: REINDEX (TABLESPACE ts) TABLE tbl; 2021-12-21 12:54:08.775 +07 [208716] LOG: terminating any other active server processes 2021-12-21 12:54:08.775 +07 [208716] LOG: all server processes terminated; reinitializing with dump: [New LWP 208725] [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Core was generated by `postgres: andrew postgres [local] REINDEX '. Program terminated with signal SIGABRT, Aborted. #0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50 50 ../sysdeps/unix/sysv/linux/raise.c: Нет такого файла или каталога. (gdb) bt #0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50 #1 0x00007feadfac7859 in __GI_abort () at abort.c:79 #2 0x000055e36b6d9ec7 in errfinish (filename=0x55e36b786e20 "xact.c", lineno=1729, funcname=0x55e36b788660 <__func__.29619> "RecordTransactionAbort") at elog.c:680 #3 0x000055e36b0d6e37 in RecordTransactionAbort (isSubXact=false) at xact.c:1729 #4 0x000055e36b0d7f64 in AbortTransaction () at xact.c:2787 #5 0x000055e36b0d88fa in AbortCurrentTransaction () at xact.c:3315 #6 0x000055e36b524f33 in PostgresMain (dbname=0x55e36d4d97b8 "postgres", username=0x55e36d4d9798 "andrew") at postgres.c:4252 #7 0x000055e36b44d1e0 in BackendRun (port=0x55e36d4d1020) at postmaster.c:4594 #8 0x000055e36b44cac5 in BackendStartup (port=0x55e36d4d1020) at postmaster.c:4322 #9 0x000055e36b448bad in ServerLoop () at postmaster.c:1802 #10 0x000055e36b448346 in PostmasterMain (argc=3, argv=0x55e36d4a84d0) at postmaster.c:1474 #11 0x000055e36b33b5ca in main (argc=3, argv=0x55e36d4a84d0) at main.c:198 Regards! On Mon, Dec 20, 2021 at 7:42 PM wenjing zeng <wjzeng2...@gmail.com> wrote: > Post GTT v63 to fixed conflicts with the latest code. > > > > Hi Andrew > > Have you found any new bugs recently? > > > > Wenjing > > > > > 2021年11月20日 01:31,wenjing <wjzeng2...@gmail.com> 写道: > > > > Andrew Bille <andrewbi...@gmail.com> 于2021年11月15日周一 下午6:34写道: > >> Thanks for the patches. The feature has become much more stable. >> However, there is another simple case that generates an error: >> Master with v61 patches >> >> CREATE GLOBAL TEMPORARY TABLE t AS SELECT 1 AS a; >> ERROR: could not open file "base/13560/t3_16384": No such file or >> directory >> > Thank you for pointing out that this part is not reasonable enough. > This issue has been fixed in v62. > Looking forward to your reply. > > > Wenjing > > > >> Andrew >> >> On Thu, Nov 11, 2021 at 3:15 PM wenjing <wjzeng2...@gmail.com> wrote: >> >>> Fixed a bug in function pg_gtt_attached_pid. >>> Looking forward to your reply. >>> >>> >>> Wenjing >>> >>> > > > <0001-gtt-v62-reademe.patch><0004-gtt-v62-regress.patch> > <0002-gtt-v62-doc.patch><0003-gtt-v62-implementation.patch> > > > > >