The monitoring of network equipment, servers and the data center itself is one of the most important tasks of the data center. Most of the relevant planning focuses on the selection of effective monitoring solutions and continuous data center network equipment management methods.
An effective monitoring solution is crucial to the maintenance of the data center network. This is because, to some extent, the data center is often left unattended for most of the time. Over the years, I have had experience working in many data centers. According to my experience, basically no one will be sent to stare at the data center to wait for problems. In fact, perhaps no one will stay in the data center for a long time. After all, data centers are cold and noisy, so network administrators prefer to spend time elsewhere. Sometimes, the administrator ’s work location may not even be in the same building as the data center. At the beginning of his career, the data center of a large insurance company where I worked was like an underground warehouse. Although the data center is well-built, only security personnel work there regularly, and others work in another building a few miles away.
Whether the administrator works in a data center building or a few miles away, an effective alert mechanism must be configured. You ca n’t just imagine someone happening to walk into the data center and notice that an upcoming problem is displayed on the console screen. This is why it is so important to ensure a good network management and monitoring solution. If not, we may only realize that something went wrong when the alarm bell sounded.
What needs to be monitored?
There are many different things in the data center that need to be monitored, so we need to formulate a large number of data center monitoring plans. Some people may think that data center monitoring is to closely observe the operation of the server, but the actual operation is much more complicated. For example, Microsoft's SystemCenterOperaTIonsManager has done a good job in monitoring Windows Servers and can also be deployed in small plans. But despite this, if our server is running a non-Windows operating system, it will not work.
In addition to server operating systems and applications, there are many other factors that we need to monitor. For example, it is also important to observe the temperature of the data center. Most servers are equipped with built-in security mechanisms. When the temperature of the server exceeds a certain value and may damage the server, it will stop the server. A good monitoring solution must be able to report the temperature of the data center environment in a timely manner, and it should also be able to alert us when the temperature of any server begins to approach a critical level.
The situation in power management is similar. When a power interruption occurs, usually the backup power will keep the server running normally within a preset time range. More complex data centers may also rely on backup generators. In any case, we must pay attention to energy outages, and we must also know how much backup power is required at any given time.
A good monitoring solution must be able to remind us to pay attention to server hardware problems, operating system failures, application failures, network hardware failures and environmental problems. This can be said to be a demanding requirement, and it is also one of the reasons why proper planning is crucial. As far as I know, no single monitoring solution can perform all these functions at the same time. Generally speaking, network architects need to purchase several monitoring solutions at the same time, and install them on the system to play an alarm role. This alert can be sent to the administrator's mobile communication device in the form of text information, or sent to the service desk in the form of an email, or to alert the administrator in the form of other alerts. The important point is that all alarms must eventually be collected in one place.
Virtualization technology complicates data center network monitoring
When we purchase monitoring solutions, we must bear in mind that there are many factors to consider, such as virtualization, which will complicate the monitoring process. For example, there are various monitoring application software on the market that can monitor server hardware failure signals. Applications such as these can monitor excessive server temperatures, SMART disk alarms, and even the failure of server cooling fans. But the problem is that if the monitoring solution does not know that it is monitoring a virtual server, it may not be able to monitor hardware issues that can potentially affect server availability.
The monitoring software should be able to identify problems with the host server hardware. However, when the host is at risk, any virtual machine running on the host may also be at risk. Therefore, when using virtual machines, we need a way to distinguish between physical servers and virtual servers, and know which virtual machines are running on which host servers. At the same time, when hardware problems occur, we must also be able to quickly transfer clients to other host servers.
The last thing to say is that management and monitoring are carried out simultaneously. Only when the management ability is in place can good monitoring be achieved, especially when employees are not working on site. For example, when the management personnel cannot reach the failed server in time to prevent the accident, if the monitoring software can inform us that there will be a major failure, how good should it be? This means that each data center server and hardware core components can be monitored and The reason why remote interaction with them is so important.
Xinxiang Mina Import & Export Co., Ltd. , https://www.mina-motor.cn