Home/ Questions/Q 4069114
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-20T16:27:37+00:00

We’re seeing this behavior randomly on newly inserted records and only on the production

  • 0

We’re seeing this behavior randomly on newly inserted records and only on the production database (Amazon’s RDS). Looks like something related to a deadlock… (no errors reported)

The affected tables are InnoDB, payment_id is auto-increment and all transaction end-times are in the order of milliseconds. All the connections are on the same server, autocommit is enabled and there is no replication (as far as Amazon’s documentation).

Connection #1:

2011-03-07 14:09:54 INSERT INTO payments SET payment_transaction = ‘XYZ’

Connection #2:

2011-03-07 14:10:06: SELECT * FROM payments WHERE payment_transaction = ‘XYZ’ LIMIT 0, 1

Response: empty

2011-03-07 14:10:06: SELECT * FROM payments ORDER BY payment_id DESC LIMIT 0, 1

Response: [payment_id] => 26242, [payment_transaction] => ABC

2011-03-07 14:50:06: SELECT * FROM payments WHERE payment_transaction = ‘XYZ’ LIMIT 0, 1

Response: [payment_id] => 26243, [payment_transaction] => XYZ

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    By adding some debug points to the application I noted that these (randomly problematic) INSERTs are actually taking around 50+ seconds to complete (in contrast to what the log was saying). So to anyone having this issue, make sure you benchmark the start and end times of the INSERT within your application environment.

    • 0