We recently upgraded from MySQL 5.7 to MySQL 8.0.35 in AWS RDS.
There is a log table that services write to and we rotate it daily, using
The table is not used by any services when we run
RENAME TABLE
command.
Later when we perform a DROP
or TRUNCATE
on this table (~10-15 GB size) the DB freezes for 5-6 minutes.
During those minutes
* DB does not report any metrics
* Clients get timeouts and disconnect
* OS metrics show significant drop in IO and CPU
Table
CREATE TABLE test_test_test
(
rowId
bigint NOT NULL DEFAULT '0',
ruId
varchar(200) NOT NULL,
sessionId
varchar(200) NOT NULL,
transSource
varchar(255) DEFAULT NULL,
dcsnType
varchar(75) DEFAULT NULL,
dcsnTypeManual
varchar(75) DEFAULT NULL,
dcsnReason
varchar(255) DEFAULT NULL,
dcsnPlcRsn
varchar(100) DEFAULT NULL,
firstDcnRule
varchar(100) DEFAULT NULL,
payload
text,
insertTime
timestamp NULL DEFAULT CURRENT_TIMESTAMP,
status
varchar(50) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci
## OS Metrics
Notice the DB metrics (bottom 3 graphs) are extrapolated (no measurements) during 07:33-07:41

DROP
or TRUNCATE
Would appreciate any theories, explanations or directions to explore.
Thanks
Asked by AlexV
(125 rep)
Feb 18, 2024, 10:57 AM
Last activity: Jul 10, 2025, 08:02 PM
Last activity: Jul 10, 2025, 08:02 PM