The architecture of 2 front-end servers and 4 Exchange back-end servers was maintained for about 6 months. During the period of recurring problems, the most prominent problem is that after Exchange 2003 dozen SP2 patches, the POP3 service often goes down, causing the client to not use POP3 service for sending and receiving mail. The problem I think may be that exchange itself is a bigger problem, and that the customer POP3 the service when the SP2 patch doesn't work. It's a very weird thing.
Other uncertain problems are basically focused on the disk IO, because the server running at the same time SMTP service and POP3 services, resulting in lengthy mail queues, server response is very slow. 4 servers have two are HP DL380 series machines. After the corresponding parameter check, found that the customer's array card cache is not open, which for applications such high IO application of the server, will severely reduce the overall performance of the server. After the array card cache was opened, we found that the queue was dropping very quickly.
Another problem is that in the Exchange 2003 database and in high availability mode, the customer's database often appears to have similar corruption. The state of the database is often found in the dirty shutdown state by ordering Eseutil to view its status.
The specific orders are as follows:
Eseutil.exe the database path. edb/mh
The Dirty shutdown database needs to be repaired, and the repair uses the following command to implement
eseutil.exe/p database path/database. edb
After the repair is complete, we may need to clear the log of the storage group before we can suspend the corresponding database
And often the above problems make the customer more sleepy, the customer because the exchange problem is not less with the client explanation. Based on Microsoft's introduction of a new generation of messaging systems, customers decided to upgrade their messaging systems to exchange 2010来 to reduce the risk of a corresponding mail server.