When I was thinking about this site’s 17th birthday, I started wondering: back in 2002, what bug might I have been working with in SQL Server?
I started reading back to the list of bugs fixed in SQL Server 2000 Service Pack 1, and it hit me…these bugs aren’t all that different than the ones we’re facing today.
Before you click on each bug, guess whether it was in SQL 2000 Service Pack 1, or last month’s SQL Server 2017 CU14. When you click on each one, you’ll see the release notes for it. (Unfortunately, the 2000 bugs all point to the SP1 page – the individual bug pages have been lost to the sands of time.) Good luck!
- CREATE for existing object with IDENTITY column causes duplicate IDENTITY values
- Statistics maintenance causes significant bottleneck on SQL Servers that use 16 or more CPUs
- Assertion failure occurs when you try to back up database in limited disk space
- TCP Timeout or login time-out error occurs when you connect to SQL Server using Integrated Authentication
- Error removing log shipping on secondary when database name has a quote
- Upgrade fails with an error when sysadmin account “sa” is renamed
- SQL Server resource DLL may fail to log to the event viewer in a cluster
- Configuration option network.enablekdcfromkrb5 now set to false instead of true
- SELECT may not return all rows if it contains a large number of values in an IN clause on a NUMERIC column
- Query results are not as expected when you run a particular query from Excel
- Complex DISTINCT or GROUP BY query can return unexpected results with parallel execution plan
- “Non-yielding” error occurs when there is a heavy use of prepared statements
- INSERT statement incorrectly conflicts with constraint
- DBCC STACKDUMP doesn’t generate dump file
- Combination of multiple EXIST, OR, and subquery clauses may give a sub-optimal plan (you know, I’m gonna go out on a limb and say that still happens)
- Incorrect record is deleted or updated when a clustered index is in descending order
- Assertion occurs when a parallel query deletes
Let me know how you did in the comments.
I did no better than random. Each one was a complete guess.
“Incorrect Record is Deleted or Updated When a Clustered Index is in Descending Order” was the one that caught my eye. It reminded me of the time when I ran our automated tests against a local database where I reversed the order of every index. I was looking for places where we assume results order without an order by and I found more than I expected.
Okay, you got me curious. What all did you find?
Hahaha, that’s such an awesome bug, too. I read that and go, “Who on earth does that?!?” SQL Server can do freakin’ anything.
Brent, that was really good exercise. My answers were maybe on 50% right (or bad). Is that mean I supposed to back to work with MS Access? 😉
I bet you’re better than most folks! When I looked at it again today, I couldn’t remember half of them, and I wrote the post, hahaha.
Wow, 13 out of 14! I got #4 wrong. Your note on #15 was a bit of a giveaway too, I reckon I would’ve guessed that as 2017 without it 😀
Got 9 right, which is no better than flipping a coin.