Content-Type: multipart/mixed; boundary="------------Q4aLw488RvmmJYjmNjreZrrJ"
Message-ID: <b12f9e22-2618-42b8-8644-88bae192c7fd@gmail.com>
Date: Tue, 28 Nov 2023 19:00:16 +0100
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Content-Language: en-US
From: David Geier <geidav.pg@gmail.com>
To: PostgreSQL-development <pgsql-hackers@postgresql.org>
Subject: Fix assertion in autovacuum worker
Archived-At: <https://www.postgresql.org/message-id/b12f9e22-2618-42b8-8644-88bae192c7fd%40gmail.com>
Precedence: bulk

This is a multi-part message in MIME format.
--------------Q4aLw488RvmmJYjmNjreZrrJ
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi hackers,

PostgreSQL hit the following assertion during error cleanup, after being 
OOM in dsa_allocate0():

void dshash_detach(dshash_table *hash_table) { 
ASSERT_NO_PARTITION_LOCKS_HELD_BY_ME(hash_table);

called from pgstat_shutdown_hook(), called from shmem_exit(), called 
from proc_exit(), called from the exception handler.

The partition locks got previously acquired by

AutoVacWorkerMain() pgstat_report_autovac() 
pgstat_get_entry_ref_locked() pgstat_get_entry_ref() 
dshash_find_or_insert() resize() resize() locks all partitions so the 
hash table can safely be resized. Then it calls dsa_allocate0(). If 
dsa_allocate0() fails to allocate, it errors out. The exception handler 
calls proc_exit() which normally calls LWLockReleaseAll() via 
AbortTransaction() but only if there's an active transaction. However, 
pgstat_report_autovac() runs before a transaction got started and hence 
LWLockReleaseAll() doesn't run before pgstat_shutdown_hook() is called.

See attached patch for an attempt to fix this issue.

-- 
David Geier
(ServiceNow)

--------------Q4aLw488RvmmJYjmNjreZrrJ
Content-Type: text/x-patch; charset=UTF-8;
 name="0001-Fix-autovacuum-cleanup-on-error.patch"
Content-Disposition: attachment;
 filename="0001-Fix-autovacuum-cleanup-on-error.patch"
Content-Transfer-Encoding: base64
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--------------Q4aLw488RvmmJYjmNjreZrrJ--