Method and device for detecting state of host system

Abstract

The invention provides a method and a device for detecting the state of a host system. The method comprises the following steps that: (a) a monitoring platform pre-establishes a secure shell (SSH) tunnel between the monitoring platform and a target host; (b) the monitoring platform detects whether the SSH tunnel is disconnected or not; (c) if the SSH tunnel is disconnected, the monitoring platform triggers the re-establishment of the SSH tunnel, repeats the step (c) if a reason for a failure in the re-establishment of the SSH tunnel is the abnormity of an SSH daemon process when the re-establishment of the SSH tunnel is failed, otherwise determines that the target host is halted; and (d) if the SSH tunnel is connected or the re-establishment of the SSH tunnel triggered by the monitoring platform is successful after the SSH is disconnected, the monitoring platform transmits a keep alive command to the target host to judge whether the target host is in a system unresponsive state or not, and re-executes the step (b) after successfully receiving the response information of the keep alive command. The system unresponsive condition can be accurately monitored in real time.

Description

technical field [0001] The invention relates to operating system monitoring technology, in particular to a method and device for detecting the state of a host computer system. Background technique [0002] In normal system monitoring, the monitoring platform of the system monitors the status of the host system, which is the most basic core monitoring index and belongs to a very important part of the monitoring system. How to discover the crash of the host in the first time, and further, how to find the abnormality of the host system in the first time is an important part of system monitoring. At present, the host system monitoring methods mainly used in the industry can be divided into Ping active monitoring method and Agent agent passive monitoring method in terms of monitoring methods. [0003] Specifically, the Ping active monitoring method is mainly: use the standard Internet Control Message Protocol (ICMP) Ping method to periodically send ICMP packets to the target hos...

AI Technical Summary

Problems solved by technology

The regular inspection method must have a certain inspection cycle. In addition, considering the load pressure of the monitoring and management server, the more servers are monitored, the longer the inspection cycle will be. Therefore, the inspection cycle must be at the minute level in general systems. of

In this way, it will not be possible to monitor the crash of the host in time

[0007] 2. For the PING active monitoring method, this method only judges whether the host is down by checking whether the network of the target host is disconnected. In actual applications, there may be situations where the network is smooth but the host system is abnormal. In this way, PING is used. The active monitoring method will not be able to judge whether the host system is abnormal

In actual host system monitoring, the judgment of host system non-response is also a very important monitoring item. When the host system does not respond, most upper-level business systems cannot provide external services normally.

[0008] 3. For the agent-based passive monitoring method, an Agent agent program needs to be installed on the monitored host. Although in theory, this method can return the system status to the monitoring and management server, but from the actual application effect , when the network communication function of the host is normal but the system does not respond, most of the system resources are consumed. At this time, the Agent on the target host cannot send information, because the Agent itself needs to consume system resources to provide external services.

Therefore, the agent-based passive monitoring method cannot realize timely monitoring of system unresponsiveness.

[0009] It can be seen that the above-mentioned existing host monitoring methods all have the problems of monitoring hysteresis and the inability to monitor the unresponsive situation of the host system

Images