Verdict:
Look to RLX to see how blade server management should be done properly. The 6G offers the best support tools with excellent monitoring capabilities and task automation.
Features, expansion, build quality and processing density are all factors to consider when choosing a blade server. But management capabilities are just as important. RLX has always offered excellent management features with its Control Tower software, but the latest 6G version takes this to a new level. There are deeper monitoring capabilities, better automated responses to problems, plus server and resource provisioning policies.
Since the introduction of the first blade server from HP in 2002, we've reviewed every new product as it's come onto the market. Our initial reaction to RLX's new management system is that nothing else comes close to this level of sophistication at this price point. The biggest drawback to this package is that it supports only RLX's own blade servers - this is no bad thing though, as we were impressed by the System 300ex and found the System 600ex good enough for a Recommended award.
The 6G is implemented as a separate Supermicro-based rack mount and installation doesn't take long. You just connect the server's gigabit Ethernet ports to a public network and the dedicated network where the blade server management controllers are located, then point a web browser at its default IP address. For testing, we used a System 600ex blade chassis equipped with five dual 3.06GHz Xeon blades with Windows Server 2003 and Linux installed on them.
The web interface opens to a simple home page with a quick graphical run-down on the number of detected servers, racks and chassis, plus monitoring status and closed and active alerts. Plenty of system information is provided under the Monitor tab and presented neatly so you can immediately see where problems are occurring. The use of agents when polling devices, as opposed to relying on just SNMP, makes the 6G far more powerful, because it gathers information more quickly and can do more with it. The Monitor view can be customised to suit as you use single components or groups and link them to monitor types. A simple example would be to view available disk space on a
ADVERTISEMENT
single blade. From the Admin tab you select the blade, create a component item for it and assign it a monitor type. RLX provides more than 30 predefined monitor types, which can be copied and modified. Individual components can be gathered into groups and have monitor items, such as a general health-check, assigned to them. A group component that appears in the monitor screen will have a status bar attached showing the number of members that are functioning correctly or have issued a warning or critical error.
The optional Provisioning module looks after automating server imaging, OS deployment and firmware or BIOS upgrades to multiple systems. The Task Manager is provided as standard and allows maintenance, troubleshooting and reporting tasks to be scheduled. A key new feature used by both modules is task chaining, where multiple tasks can be assigned to a single job. You could, for example, automatically update servers with new firmware and reboot them. But task chains can also be fired up in response to a fault or error detected by the monitoring component. A number of predefined tasks are included, and we found creating new tasks was made more difficult by the lack of detailed documentation. The tasks are placed in a chain in the order they're to be processed, but there are limitations as task failures aren't catered for, so the chain will fail.
The Manage tab delivers some of the best server blade and chassis monitoring facilities we've yet seen, and the use of digital pictures rather than simple graphics makes it look so much slicker. You can view the front and rear of the chassis and easily see the status of individual components. Even the LCD panels on each server and management blade are revealed. Selecting the latter component takes you straight over to RLX's ActiveConfig Java-based remote interface for closer control of the chassis components. The levels of operational information provided by 6G are impressive. Graphics are provided for all chassis power supplies, voltages, fans and temperatures, and selecting a server blade gives you full control over it. The benefits of the new agent become apparent as you can also view OS-related information such as the network configuration, running processes and installed software.
Having seen the blade management offerings from virtually all vendors over the past couple of years, we can safely say the Control Tower 6G is the best yet. The sheer number of features on offer does actually make it more complex to use and only RLX users need apply. But for automating general support tasks and managing large blade server installations this sets a very high standard.
By Dave Mitchell
SPECIFICATIONS:
Supermicro SC812 1U rack-mount chassis; Supermicro X5DPR-IG2+ motherboard; 2GHz Intel Xeon; Intel E7501 chipset; 1GB PC2100 ECC SDRAM expandable to 12GB; 3ware Escalade 8006-2LP SATA/150 dual-channel PCI RAID card, supports RAID0, 1 and JBODs; 2 x 80GB Maxtor DiamondMax Plus 9 SATA/150 hard disks; 2 x Intel gigabit Ethernet; Linux kernel. Options: Provisioning Manager, £105 per server; Policy Manager, £210 per server.