Re: [Proposal] Global temporary tables

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 <wjzeng2012@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 <wjzeng2012@gmail.com> 写道：

Andrew Bille <andrewbille@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 <wjzeng2012@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>