每隔几个月,大约30分钟,我看到我的SQL Server CPU出现了大约80%的使用率。
我终于在tempdb上运行此查询的行为中发现了它:
SET NOCOUNT ON;
DECLARE @previous_collection_time datetime;
DECLARE @previous_request_count bigint;
DECLARE @current_collection_time datetime;
DECLARE @current_request_count bigint;
DECLARE @batch_requests_per_sec bigint;
DECLARE @interval_sec bigint;
-- Get the previous snapshot's time and batch request count
SELECT TOP 1 @previous_collection_time = collection_time, @previous_request_count = request_count
FROM #am_request_count
ORDER BY collection_time DESC;
-- Get the current total time and batch request count
SET @current_collection_time = GETDATE();
SELECT @current_request_count = cntr_value
FROM sys.sysperfinfo
WHERE counter_name = 'Batch Requests/sec' COLLATE Latin1_General_BIN;
SET @interval_sec =
-- Avoid divide-by-zero
CASE
WHEN DATEDIFF (second, @previous_collection_time, @current_collection_time) = 0 THEN 1
ELSE DATEDIFF (second, @previous_collection_time, @current_collection_time)
END;
-- Calc the Batch Requests/sec rate for the just-completed time interval.
SET @batch_requests_per_sec = (@current_request_count - @previous_request_count) / @interval_sec;
-- Save off current batch count
INSERT INTO #am_request_count (collection_time, request_count)
VALUES (@current_collection_time, @current_request_count);
-- Return the batch requests/sec rate for the just-completed time interval.
SELECT ISNULL (@batch_requests_per_sec, 0) AS batch_requests_per_sec;
-- Get rid of all but the most recent snapshot's data
DELETE FROM #am_request_count WHERE collection_time < @current_collection_time;
不是SQL Server的专家,有谁知道发生了什么?我怎样才能阻止这个/安排这个?