|< Day Day Up >|| |
Other vendors have put out products similar to CIM, although they are not as comprehensive. These include Dell OpenManage, HP TopTools, and IBM Netfinity Manager. CIM can interface with some of these tools and is able to monitor non-Compaq drives. Similarly, some of these vendor tools can pick up some of the information contained in CIM, but none of them does a complete job and most founder when it comes to monitoring anything at all on a clone server. This is particularly a problem outside of the United States where local hardware manufacturers hold a much larger market share compared with these big original equipment manufacturers (OEMs). Most of these OEM tools use either the Simple Network Management Protocol (SNMP) or the Desktop Management Interface (DMI). SNMP is a widely used network monitoring and control protocol. SNMP agents pass data about devices to the console that is monitoring them. These agents return data contained in a Management Information Base (MIB), a data structure defining what can be obtained from each device and the ways in which it can be controlled. Similarly, DMI is a management system that can monitor hardware and software components from a central console. Agents are used to gather data when queries are made. CIM uses SNMP, and Dell OpenManage uses DMI.
To avoid the frailties of SMART, Gibson Research's SpinRite tests drives by reading and rewriting data to the entire surface of a disk, removing any bad sectors as it goes. It issues alerts of potential drive failures and offers a nice set of recovery tools, too. This fine utility, though, has one drawback — it only works on FAT partitions, limiting it to Windows 9X and NT partitions running FAT. According to the Gibson Web site, a version will eventually be coming out that addresses NTFS and other file systems. It is hoped that when this version appears it will also be networkable, enabling companies to load it on their networks to monitor the health of every single server and desktop disk. Now that would be quite a product!
Though SMART drives have certainly failed to live up to early expectations, products such as CIM hold out some hope that the technology will not, in the end, prove less than useful. SpinRite offers a simpler and more effective solution to hard drive efficiency, at least for the individual machine. Some, however, might point to hot swapabilty as the answer. In theory, with a hotswappable RAID array, when one drive goes down users can just replace it on the fly and they are up and running again. In the real world, though, it does not quite work out this way.
Say, for example, a Web server drive has been steadily but slowly deteriorating. On Friday, everything is fine, but first thing Monday morning it is discovered that one of the RAID drives is down. Due to RAID redundancy, the Web server has continued to function since the drive failed on Saturday morning, but performance has suffered badly. User complaints are piling up, and the boss is mad because the Web site hits are taking a beating. Even when the drive is quickly replaced, the company's troubles are far from over. Even after everyone is informed that the new drive is now online, the complaints do not let up and site usage bottoms out. Hundreds of thousands of dollars in business just went to the competition. What happened? Simply put, it takes many hours for a RAID array to rebuild after a new drive is installed. The more prudent approach, therefore, is to know about drive failures well in advance so appropriate action can be taken. Had the alert been received the week before, the faulty drive could have rebuilt during off peak hours, and the company could have continued to make a bundle in online sales.
|< Day Day Up >|| |