Verifying that data_engine is deleting old raw and historic data properly from the database
Article ID: 34940
Updated On:
Products
Issue/Introduction
How can we verify that the data_engine is deleting old raw and historic data properly from the database?
Environment
Release: UIM 9.x
Resolution
-
Access your SQL server's desktop
Log into a SQL Management Studio console
Select the appropriate database from the dropdown in the toolbar. This will usually be NimsoftSLM for older installations or CA_UIM for newer.
Paste the following query into a New Query window and click the Execute button/****** Script for finding old data which the data_engine failed to summarize ******/
DECLARE @sql NVARCHAR(256),@tbl_char VARCHAR(4),@count INT,@count_out INT,@index INT = 1,@tbl_num INT,@raw_days INT,@hist_days INT,@daily_days INT,@rows_found TINYINT = 0,@rows_out NVARCHAR(128),@params NVARCHAR(256) = N'@rowcountOUT INT, @count_out INT OUTPUT';
SET @raw_days = (SELECT RawAge FROM tbn_de_DataMaintConfig) + 2; -- Days where the data_engine should delete raw data
SET @hist_days = (SELECT HistoryAge FROM tbn_de_DataMaintConfig) + 2; -- Days where the data_engine should delete hourly data
SET @daily_days = (SELECT DailyAvgAge FROM tbn_de_DataMaintConfig) + 2; -- Days where the data_engine should delete daily data
-- Set @tbl_num to be the number of raw tables (a count of the S_QOS_DEFINTION table)
SET @sql = N'SELECT @count_out = COUNT(*) FROM S_QOS_DEFINITION';
EXEC sys.sp_executesql @sql, @params, @rowcountOUT = @count, @count_out = @rows_out OUTPUT;
SET @tbl_num = CAST(@rows_out AS INT);
-- Search the RN tables for data that should have been removed
WHILE (@index <= @tbl_num) BEGIN
SET @tbl_char = RIGHT('0000' + CONVERT(VARCHAR(4), @index), 4);
SET @sql = N'SELECT @count_out = COUNT(*) FROM RN_QOS_DATA_' + @tbl_char + ' WHERE sampletime < (GETDATE() - ' + CAST(@raw_days AS VARCHAR(5)) + ')';
-- Make sure the table exists before running SQL against it ...
IF EXISTS(SELECT name FROM sysobjects WHERE name = N'RN_QOS_DATA_'+ @tbl_char and xtype ='U')
EXEC sys.sp_executesql @sql, @params, @rowcountOUT = @count, @count_out = @rows_out OUTPUT;
-- Print the output of old raw data
IF (CAST(@rows_out AS INT) > 0) BEGIN
PRINT 'RN_QOS_DATA_' + @tbl_char + ' contains ' + @rows_out + ' samples which are older than ' + CAST(@raw_days AS VARCHAR(4)) + ' days old.';
SET @rows_found = 1;
END
SET @index = @index + 2;
END
IF (@rows_found = 0)
PRINT 'No unmaintained data found in RN tables';
-- Reset our index variable to traverse HN tables
SET @index = 1;
SET @rows_found = 0;
-- Search the HN tables for data that should have been removed
WHILE (@index <= @tbl_num) BEGIN
SET @tbl_char = RIGHT('0000' + CONVERT(VARCHAR(4), @index), 4);
SET @sql = N'SELECT @count_out = COUNT(*) FROM [HN_QOS_DATA_' + @tbl_char + '] WHERE sampletime < (GETDATE() - ' + CAST(@hist_days AS VARCHAR(5)) + ')';
-- Make sure the table exists before running SQL against it ...
IF EXISTS(SELECT name FROM sysobjects WHERE name = N'HN_QOS_DATA_' + @tbl_char and xtype = 'U')
EXEC sys.sp_executesql @sql, @params, @rowcountOUT = @count, @count_out = @rows_out OUTPUT;
-- Print the output of old historic data
IF (CAST(@rows_out AS INT) > 0) BEGIN
PRINT 'HN_QOS_DATA_' + @tbl_char + ' contains ' + @rows_out + ' samples which are older than ' + CAST(@hist_days AS VARCHAR(4)) + ' days old.';
SET @rows_found = 1;
END
SET @index = @index + 2;
END
IF (@rows_found = 0)
PRINT 'No unmaintained data found in HN tables';
-- Reset our index variable to traverse DN tables
SET @index = 1;
SET @rows_found = 0;
-- Search the DN tables for data that should have been removed
WHILE (@index <= @tbl_num) BEGIN
SET @tbl_char = RIGHT('0000' + CONVERT(VARCHAR(4), @index), 4);
SET @sql = N'SELECT @count_out = COUNT(*) FROM [DN_QOS_DATA_' + @tbl_char + '] WHERE sampletime < (GETDATE() - ' + CAST(@daily_days AS VARCHAR(5)) + ')';
-- Make sure the table exists before running SQL against it ...
IF EXISTS(SELECT name FROM sysobjects WHERE name = N'DN_QOS_DATA_' + @tbl_char and xtype = 'U')
EXEC sys.sp_executesql @sql, @params, @rowcountOUT = @count, @count_out = @rows_out OUTPUT;
-- Print the output of old historic data
IF (CAST(@rows_out AS INT) > 0) BEGIN
PRINT 'DN_QOS_DATA_' + @tbl_char + ' contains ' + @rows_out + ' samples which are older than ' + CAST(@hist_days AS VARCHAR(4)) + ' days old.';
SET @rows_found = 1;
END
SET @index = @index + 2;
END
IF (@rows_found = 0)
PRINT 'No unmaintained data found in DN tables';
/****** End SQL Script ******/
If the output from the script shows as follows then data_engine is honoring the retention policies configured in data_engine:
No unmaintained data found in RN tables
No unmaintained data found in HN tables
No unmaintained data found in DN tables
If, however you receive output that looks like the following, then a support case may need to be opened to fix the issue
RN_QOS_DATA_0001 contains 416 samples which are older than 30 days old.
RN_QOS_DATA_0007 contains 2524 samples which are older than 30 days old.
RN_QOS_DATA_0008 contains 415 samples which are older than 30 days old.
Please also note that this query can take 10 - 15 minutes or more to run on a large database.
Additional Information
Article title: data_engine best practices
https://ca-broadcom.wolkenservicedesk.com/external/article?articleId=33592
Attachments
Feedback
