Anyone who looks at raw Dynamics GP tables is familiar with this column DEX_ROW_ID that occurs on most of them. It is an identity column, it auto increments an integer for each new record inserted. In GP it is an of integer data type, this means if the column was seeded with an initial value of -2,147,483,648 it would increment up to 2,147,483,647 before running out of numbers to use. GP starts and zero as a seed so this halves the quantity of numbers available.
Today I started getting notifications from GP of an error that users were experiencing. It was essentially the following error when trying to use a product creation wizard add-in that I created in GP.
Message: Arithmetic overflow error converting IDENTITY to data type int.
Arithmetic overflow occurred.
AS it turns out this is an issue I had expected to bite one day and never did anything about it. I feel guilty, however on the plus side I found the issue very quickly, it being related to the third table I checked for issues. You may want to use the script further down to find the problem table if you are reading this while fire fighting a production issue that is similar in nature.
The error indicates that the IDENTITY column of a table is trying to use a value over that which the data type (int) supports. In plain English this means the DEX_ROW_ID of one of our tables is trying to use a number over 2,147,483,647 and it can’t as it is not allowed as that is bigger than an integer. This does not mean there will be that many records in the table, as some may have been deleted since they were created. DBCC CHYECKIDENT in SQL can help find the next identity number about to be used.
In this case there are overnight processes that manage the 1.5 million selling price points, triggered from changes in item standard cost and exchange rates. This over the years builds up to a lot of inserts and deletes of price (currency) records. It would seem last night that process broke the IV00105 table, the currency price table. – later I found an error report from that process of that issue in my in box. I would have found it after my morning coffee, had this not all kicked off first.
So what to do?
We have a large number of users hammering GP all day so downtime is not desirable, so thinking on my feet I looked to see what the min and max values were in GP for DEX_ROW_ID of that table.
select MAX(DEX_ROW_ID), MIN(DEX_ROW_ID) from iv00105
I was originally running this, intending to reseed back to zero as my assumption was there would be, as there is, a gap that would be big enough to get though the day.
However look again 2147403452 is not 2147483647, the maximum integer value.
I checked what the actual next number the column would use would be with the DBCC CHECKIDENT command:
DBCC CHECKIDENT (IV00105)
Indeed this verified the identity column was currently sat at 2147483647. This is helpful, some identity numbers must have been used up but records never created, something in itself that needs checking out later, but for now it gave me the option of simply reseeding the identity column to the last used number, 2147403452. Thus new records could then reuse the gap up to the ceiling limit. Executing the following SQL reset the seed on the table and let the users carry on.
DBCC CHECKIDENT (IV00105,RESEED,2147403453)
Problem solved for today -as there are now 2147483647 - 2147403452 = 80,195 records for use today, before again reaching the limit of the column. That is plenty enough for one day, tonight I will renumber the DEX_ROW_ID column to take out any gaps, compressing it to take minimum number of values up, after which I will reseed to the new highest value. That should keep us going for another ten years…
Detecting DEX_ROW_ID identity columns near to the limit
Lastly I was bothered that there may be other tables close to the limit, for example I know the IV00108 gets a lot of activity too. So I looked for a monitoring solution as prevention is better than cure.
I found Shaun J Stuart has a script similar to others I’ve seen on stack overflow and others from Google sources, this would seem to suit my purpose!
I’ve added a parameter for setting the threshold and reproduced it here in case that source site ever goes down:
DECLARE @SchemaName sysname
DECLARE @TableName sysname
DECLARE @ColumnName sysname
DECLARE @threshold DECIMAL(3,2);
SET @threshold = .95;
DECLARE @Name sysname
DECLARE DBNameCursor CURSOR
WHERE source_database_id IS NULL /* no database snapshots */
AND is_read_only = 0 /* no read-only dbs (log shipping) */
AND database_id <> 2 /* skip tempdb */
AND state = 0 /* online databases */
ORDER BY name;
CREATE TABLE #IdentityCheck
PK INT IDENTITY(1, 1)
FETCH NEXT FROM DBNameCursor INTO @Name
WHILE @@fetch_status = 0
EXEC('USE [' + @Name + ']; INSERT INTO #IdentityCheck
SELECT ''' + @Name + ''' as DatabaseName
,SCHEMA_NAME(o.SCHEMA_ID) AS SchemaName
,OBJECT_NAME(o.OBJECT_ID) AS TableName
,c.[name] AS ColumnName
,t.name AS ColumnType
,CAST(last_value AS bigint) AS MaxValue
FROM sys.identity_columns c
WHERE c.is_identity = 1
AND t.system_type_id = c.system_type_id
AND o.object_id = c.object_id
AND o.type = ''u''
AND t.is_user_defined = 0 /* only look at system defined types */
AND t.name IN (''int'', ''bigint'', ''smallint'', ''tinyint'')')
FETCH NEXT FROM DBNameCursor INTO @Name
SELECT 'Table [' + SchemaName + '].[' + TableName + '], column ['
+ ColumnName + '] in database [' + DatabaseName
+ '] is an identity column that has reached over 90% of the maximum value for that datatype.' AS Details
,'' AS ChangeRequirements
WHERE 1 = CASE WHEN ColumnType = 'bigint'
AND (CONVERT(DECIMAL(19, 0), MaxValue)
/ 9223372036854775807.0) > @threshold THEN 1
WHEN ColumnType = 'int'
AND (CONVERT(DECIMAL(10, 0), MaxValue)
/ 2147483648.0) > @threshold THEN 1
WHEN ColumnType = 'smallint'
AND (CONVERT(DECIMAL(5, 0), MaxValue) / 32767.0) > @threshold
WHEN ColumnType = 'tinyint'
AND (CONVERT(DECIMAL(3, 0), MaxValue) / 255.0) > @threshold
DROP TABLE #IdentityCheck
--Read more: http://shaunjstuart.com/archive/2013/05/identifying-identity-columns-approaching-their-limit/#ixzz4PPdawizh
--Under Creative Commons License: Attribution Non-Commercial No Derivatives
Running the above gives results as two tables, the table from the production and the same from test (restored from production) companies. These are IV00105, the one causing me a problem today.
So it can be seen that this script should be added into our SQL alerting system, to email this result set when ever there are any results to the query. Thus providing an early warning of this issue happening again in the future.
Compressing the identity column
For this table I used the following script, I do so because know there are NO
- cascading deletes
- any change tracking
- table relationships
I would warn you to be very careful on how you decide to tackle this and think about what impact ripping the records out and putting them back again may have before doing it. There are some different approaches that could be used, as always check on a test environment first! I cannot recommend you use the same script for your system as I don’t know your install and how it would interplay with it.
--Create copy of data
select * into #temp From IV00105
--Empty exiting table
truncate table IV00105
--Reseed identity to 1
DBCC CHECKIDENT (IV00105, RESEED, 1)
--Put the records back
insert into IV00105
DROP table #temp
--Check it looks good
SELECT TOP (100) * FROM IV00105 ORDER BY DEX_ROW_ID ASC
SELECT TOP (100) * FROM IV00105 ORDER BY DEX_ROW_ID DESC
--Seed look good too compared to max record?
DBCC CHECKIDENT (IV00105)
If this helped let me know with a comment…