Discuss ASP suddenly inaccessible due to deadlock or process congestion

Symptom:
This is often the case when I install a counting system or a reverse link system on dc9.cn. The phenomenon is that your space Suddenly cannot be accessed. if you carefully observe that only ASP cannot be accessed, but static pages can be accessed, ftp access is also normal, and it seems that IIS is suspended.

Ask:
First, I called the Unlimited customer service. They had a good attitude and said that the server may be inaccessible due to process congestion. They helped me reclaim the server resources for access.

Research:
What has caused sudden deadlocks or process blocking on my server? After studying for a long time, I finally concluded that ASP was suddenly inaccessible due to the large access volume of the statistics system and the large number of users who queried and updated the database.

The reason may be:
Every time a database is queried and updated, the requested resources are not closed or released in time, leading to process congestion.
User Q accesses Table A (Lock table A) and then table B. In this case, user W accesses table B (Lock table B) and then tries to access Table, in this case, user Q has locked table B and must wait for user W to release table B to continue. Similarly, user W has to wait for user Q to release Table A to continue, resulting in a deadlock.
It's like I locked my house and ran to your house with the key. Then you locked me in your house and attempted to enter my house. In this way, you cannot get in when I try again, so I will wait for nothing.
Or user a reads a record E and then modifies the record e, but it is not locked. In this case, when user B modifies the record e, it is easy to cause a deadlock. This is mainly because the database locking type is improperly set, resulting in severe resource competition, resulting in deadlocks.
For details, see:
How can this problem be solved?
1. In my opinionProgramPay attention to the habit of using statements such as XXX. Close: Set xxx = nothing to release server resources in time. At the beginning, my sxna1.3 failed to release server resources in time, resulting in many servers suspended for death. Later, I made some basic corrections.
2. I think it is difficult to use adlockpessimisticin statements on the Data Warehouse locks. Because the default locking method of conn.exe cute is adlockreadonly, it is easy to cause deadlocks when a large number of users access the query and update database at the same time.
3. Use MSMQ to control ASP processes and solve process deadlocks. There are a lot of tutorials on the network, so I will not go into details.
In short, this problem should be caused by a program bug.

My solution is the most enjoyable solution. Instead of using station statistics, I use off-site statistics. the free traffic statistics outside the site are also good, and the reverse link can also be seen clearly, but you just need to click to see it.