When deploying a mailbox server in an enterprise, we are pursuing a primary goal: High Availability of the Exchange Mailbox server. In general, it is how to reduce the downtime of the mailbox server. It is our mailbox administrator's responsibility to ensure that enterprise email servers can run without faults. However, the moon is overcast and clear, and the sky is unpredictable, and the mailbox server will also be ill. Therefore, when deploying the mailbox server, we also need to evaluate the availability of the mailbox server to see what risks exist, which may reduce the availability of the mailbox server.
There is no quantitative indicator in the evaluation of mailbox server availability, and more is to summarize previous experiences. Below I will list some of my experiences in this area, which will be helpful to you in the future. In general, we can assess the risks of high availability of the mailbox server from five aspects.
1. Environmental Risks.
Because it takes 365 days a year for an email server to run on a daily basis, it can be regarded as a delicate device with high requirements for the external environment. When evaluating the availability of the mailbox server, the author often starts from a relatively simple environment problem.
Check the power supply status if necessary. Continuous Power Supply and stable voltage are the basic requirements for mailbox server availability. Unfortunately, few companies can do this, especially in terms of voltage. It is difficult to ensure its voltage regulation. Therefore, when deploying an enterprise mailbox server, I suggest you deploy a UPS with a voltage regulation function in terms of power supply. This device can ensure a stable current for the mailbox server. When an unexpected power failure occurs, it can also be mentioned as a protection function.
If the data center has any disaster prevention measures. My company has many servers. For this reason, the company has a dedicated office to store servers. This storage server space is independent from our office. That is to say, no one can store all kinds of servers in the room. Because people walk in and out, there will be a lot of dust. What's more, it's inevitable. Therefore, the author suggests that enterprises, whether mailbox servers or other server applications, should be able to divide an independent space to store them. In addition, cameras or alarms can be configured to prevent servers from being stolen. Therefore, whether the data center's disaster prevention measures are appropriate is also an important factor affecting the availability of email servers.
There are also considerations for this environmental problem, such as dust, temperature, humidity, line layout, and so on. When enterprises have strict high availability requirements on mailbox servers, these issues cannot be ignored.
Second, hardware risks.
When deploying the mailbox server, hardware also affects the availability of the mailbox server.
Whether redundant power supply exists. Because the mailbox server requires 365 days of continuous operation every year, it is a great test of power supply. According to the author's understanding, even if the server's power supply has taken into account this problem and adopted good materials, the server often becomes a machine due to power failure. Therefore, to reduce the impact of server faults caused by server power supply problems, the recommended method is to use redundant power supply. For example, when I deploy an email server to a customer, I suggest that enterprises purchase servers with redundant power when purchasing them, that is, servers with two power supplies. In addition, when a power supply fails, it needs to be replaced immediately. In this case, when a power supply problem occurs, another redundant power supply can take over immediately to prevent power failure and improve the availability of the mailbox server.
If the disk space is not properly planned, the availability of the mailbox server will be reduced. The most common cause is that the application stops responding due to insufficient disk space. Therefore, in the disk space planning, on the one hand, you need to purchase a large enough hard disk, but now the hard disk price is not expensive. On the other hand, you can consider using the disk quota. In my actual work, I found that many employees do not have the habit of clearing their mailboxes on a regular basis. They have been working for ten years in the enterprise, and their emails may be retained for ten years on the mailbox server, this is a waste of disk space. In addition, some employees also use the mailbox server as a file backup server, and simply put some important files in the mailbox server. In this case, you can set a disk quota for each user during disk planning to reduce the pressure on disk space. Finally, when planning a disk, you need to set a disk space alarm. When the disk space is insufficient, such as the usage has reached80%. So that the administrator can have enough time to deal with this issue, whether to add a hard disk or clear files in the hard disk. In this case, you can try again until the hard disk space is exhausted.
Therefore, the mailbox server's Hardware settings also affect its availability. In terms of hardware configuration, redundancy is an important measure to ensure high availability.
Third: Risks of application design.
The availability of the mailbox server is also directly related to the application design. Here, the application includes three parts:Operating SystemThe second is the mailbox server itself, and the third is some third-party applications.
When deploying the Exchange2007 server, the preferred operating system is of course Microsoft's server operating system. Although the stability of Microsoft's servers remains to be improved, so far, we have no choice but to deploy Exchange Mailbox servers on operating systems other than Microsoft servers, for example, a more stable Unix server operating system. Since we cannot change it, we can only accept it. When deploying an operating system, we need to consider its security. Such as applying the latest patch and disabling unnecessary services and ports. To improve the availability of the Exchange Mailbox server, you must first improve the operating system stability.
The design of the Exchange2007 mailbox server is very good. In terms of stability, reliability, and security, it has an essential improvement over 2003.
In addition to the availability of the operating system and email server, you must note the risks caused by some third-party applications. The most common example is to deploy anti-virus software on the mailbox server. When the current anti-virus software is exposed to the dark screen, we need to be more cautious when choosing anti-virus software. You must be aware that the anti-virus software cannot kill viruses. Instead, the operating system or the mailbox server system file is killed as a virus, which causes these systems to crash. This is not worth the candle.
In addition to software design, the stability of applications is most important to the configuration of these software. Even if the application software is designed best, its availability will be greatly compromised if the configuration is unreasonable. In subsequent articles, I will elaborate on the content in this area.
Fourth: process risks.
When deploying the mailbox server, we always emphasize the process of mailbox server maintenance, that is, we need to develop a certain process to standardize the operations of the Administrator, so as to avoid expanding the problem to an unmanageable level. In general, we need to set a standard process for maintenance of any mailbox server. In daily maintenance, we should strictly follow this process to avoid expansion of the problem.
For example, I have a lot of standard procedures for mailbox server maintenance. For example, you may need to migrate your emails from one email server to another. The first step and the second step are clearly defined. With this process specification, even if you have never been in touch with the Exchange Mailbox server before, you can complete the task step by step as long as you strictly follow this process, it may be slower. In this case, we can avoid irreparable losses due to temporary negligence.
For some common maintenance tasks, such as user account suggestions, Disk Quota Adjustment, and Spam Policy Formulation and handling, a detailed job process is required. In the maintenance process, follow the process strictly. Only in this way can the risks caused by the process be reduced.
Fifth, personnel allocation risks.
Staffing is also an important factor that directly affects mailbox server availability. Generally, a full range of high-level mailbox administrators is an important guarantee for the high availability of mailbox servers. However, this requires the enterprise to spend a lot of manpower costs. Therefore, enterprises need to achieve a balance between the high availability and staffing of the mailbox server. If the enterprise does not require the availability of the mailbox server, for example, to allow the mailbox server to have a one-day failure time, You Can slightly change the personnel configuration, this is because you can seek help from the outside in case of a difficult fault. However, for example, if an enterprise requires a high availability of the mailbox server and allows a maximum of two hours or even shorter failure times, it should consider reducing the other risks mentioned above, some high-level talents need to be considered. Because these technicians can respond in time not only when there is a problem with the mailbox server. You can even improve the monitoring of the running status of the mailbox server to solve some problems in the bud. This is the main role of staffing.
This article is transferred from: http://blog.csdn.net/onlyzhangqin/archive/2008/10/10/3048328.aspx