Commit 680d5405 authored by Tom Lane's avatar Tom Lane

Avoid unnecessary failure in SELECT concurrent with ALTER NO INHERIT.

If a query against an inheritance tree runs concurrently with an ALTER
TABLE that's disinheriting one of the tree members, it's possible to get
a "could not find inherited attribute" error because after obtaining lock
on the removed member, make_inh_translation_list sees that its columns
have attinhcount=0 and decides they aren't the columns it's looking for.

An ideal fix, perhaps, would avoid including such a just-removed member
table in the query at all; but there seems no way to accomplish that
without adding expensive catalog rechecks or creating a likelihood of
deadlocks.  Instead, let's just drop the check on attinhcount.  In this
way, a query that's included a just-disinherited child will still
succeed, which is not a completely unreasonable behavior.

This problem has existed for a long time, so back-patch to all supported
branches.  Also add an isolation test verifying related behaviors.

Patch by me; the new isolation test is based on Kyotaro Horiguchi's work.

Discussion: https://postgr.es/m/20170626.174612.23936762.horiguchi.kyotaro@lab.ntt.co.jp
parent 90947674
......@@ -1832,7 +1832,7 @@ make_inh_translation_list(Relation oldrelation, Relation newrelation,
*/
if (old_attno < newnatts &&
(att = TupleDescAttr(new_tupdesc, old_attno)) != NULL &&
!att->attisdropped && att->attinhcount != 0 &&
!att->attisdropped &&
strcmp(attname, NameStr(att->attname)) == 0)
new_attno = old_attno;
else
......@@ -1840,7 +1840,7 @@ make_inh_translation_list(Relation oldrelation, Relation newrelation,
for (new_attno = 0; new_attno < newnatts; new_attno++)
{
att = TupleDescAttr(new_tupdesc, new_attno);
if (!att->attisdropped && att->attinhcount != 0 &&
if (!att->attisdropped &&
strcmp(attname, NameStr(att->attname)) == 0)
break;
}
......
Parsed test spec with 2 sessions
starting permutation: s1b s1delc1 s2sel s1c s2sel
step s1b: BEGIN;
step s1delc1: ALTER TABLE c1 NO INHERIT p;
step s2sel: SELECT SUM(a) FROM p; <waiting ...>
step s1c: COMMIT;
step s2sel: <... completed>
sum
11
step s2sel: SELECT SUM(a) FROM p;
sum
1
starting permutation: s1b s1delc1 s1addc2 s2sel s1c s2sel
step s1b: BEGIN;
step s1delc1: ALTER TABLE c1 NO INHERIT p;
step s1addc2: ALTER TABLE c2 INHERIT p;
step s2sel: SELECT SUM(a) FROM p; <waiting ...>
step s1c: COMMIT;
step s2sel: <... completed>
sum
11
step s2sel: SELECT SUM(a) FROM p;
sum
101
starting permutation: s1b s1dropc1 s2sel s1c s2sel
step s1b: BEGIN;
step s1dropc1: DROP TABLE c1;
step s2sel: SELECT SUM(a) FROM p; <waiting ...>
step s1c: COMMIT;
step s2sel: <... completed>
sum
1
step s2sel: SELECT SUM(a) FROM p;
sum
1
starting permutation: s1b s1delc1 s1modc1a s2sel s1c s2sel
step s1b: BEGIN;
step s1delc1: ALTER TABLE c1 NO INHERIT p;
step s1modc1a: ALTER TABLE c1 ALTER COLUMN a TYPE float;
step s2sel: SELECT SUM(a) FROM p; <waiting ...>
step s1c: COMMIT;
step s2sel: <... completed>
error in steps s1c s2sel: ERROR: attribute "a" of relation "c1" does not match parent's type
step s2sel: SELECT SUM(a) FROM p;
sum
1
......@@ -59,6 +59,7 @@ test: multiple-cic
test: alter-table-1
test: alter-table-2
test: alter-table-3
test: alter-table-4
test: create-trigger
test: sequence-ddl
test: async-notify
......
# ALTER TABLE - Add and remove inheritance with concurrent reads
setup
{
CREATE TABLE p (a integer);
INSERT INTO p VALUES(1);
CREATE TABLE c1 () INHERITS (p);
INSERT INTO c1 VALUES(10);
CREATE TABLE c2 (a integer);
INSERT INTO c2 VALUES(100);
}
teardown
{
DROP TABLE IF EXISTS c1, c2, p;
}
session "s1"
step "s1b" { BEGIN; }
step "s1delc1" { ALTER TABLE c1 NO INHERIT p; }
step "s1modc1a" { ALTER TABLE c1 ALTER COLUMN a TYPE float; }
step "s1addc2" { ALTER TABLE c2 INHERIT p; }
step "s1dropc1" { DROP TABLE c1; }
step "s1c" { COMMIT; }
session "s2"
step "s2sel" { SELECT SUM(a) FROM p; }
# NO INHERIT will not be visible to concurrent select,
# since we identify children before locking them
permutation "s1b" "s1delc1" "s2sel" "s1c" "s2sel"
# adding inheritance likewise is not seen if s1 commits after s2 locks p
permutation "s1b" "s1delc1" "s1addc2" "s2sel" "s1c" "s2sel"
# but we do cope with DROP on a child table
permutation "s1b" "s1dropc1" "s2sel" "s1c" "s2sel"
# this case currently results in an error; doesn't seem worth preventing
permutation "s1b" "s1delc1" "s1modc1a" "s2sel" "s1c" "s2sel"
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment