It's good practice to build redundancy into your VIOS configuration so that client LPARs can use multiple VIOS, but how would you know if one—or both—had failed?
There's a popular misconception in the industry that Virtual I/O Server (VIOS) is just a slimmed down version of AIX. If you come from an AIX background, you will find the command line interface (CLI) to be much more familiar than an IBM i administrator would, for example, but the similarities stop there.
VIOS is different, and it's different enough for users of IBM Power Systems to engage with IBM Business Partners and other specialized resources to access the necessary configuration skills required for the initial configuration setup. While there's nothing wrong with this, when the business partner or specialist departs, the expertise goes too, leaving a black hole where there once was VIOS knowledge.
Underlining this point is this stat from the recent IBM i Marketplace Survey: a staggering 78 percent of respondents commented that they either don't use or don't know if they use VIOS on their IBM Power Systems server.
How Can I Monitor VIOS?
A number of command line utilities are provided to help you obtain performance-related information from your VIOS partitions:
- Topas—Much the same as its AIX equivalent, this command provides information such as disk space usage, as well as process, memory, and network adapter utilization figures. This command can be used in standard mode or nmon mode.
- Nmon—This open-source tool provides more than 25 performance-related metrics and can be used in either online or capture-style recording mode.
- Vmstat—This command reports processes, paging, memory usage, I/O, and CPU activity statistics.
- Svmon—The svmon command displays information about the current state of memory.
- Sysstat—The sysstat command provides information on the amount of time since the last system startup, the number of users logged in, and number of processes running.
- Viostat—This command is used for system input and output device loading and showing the time the physical disks are active in relation to their average transfer rates.
- Lsmap—This command details the mapping between virtual host adapters and the physical devices with which they are associated.
- Lsvg—The lsvg shows high-level information related to volume groups.
- Lslv—This command displays the characteristics and status of logical volumes.
- Lspv—This command lists pertinent information about physical volume group volumes.
- Seastat—The seastat command provides statistics on shared Ethernet adapters (SEA) managed by the VIOS partition.
There are numerous switches and flags on the above commands. Depending on what you're looking to do with the information provided, some of these commands have the ability to provide both interactive, regularly updated, real-time statistics and verbose data collections, preferred for historical performance analysis work.
How Are My VIOS Partitions Performing?
With VIOS version 220.127.116.11 comes a Performance Advisor. Designed to be installed quickly and without fuss, and consuming negligible resources, the advisor runs and collates VIOS partition performance data, analyzing the results before finally producing an easy-to-read XML file containing an at-a-glance health-check report.
The report covers many elements associated with the general performance of disk adapters, shared processors, and memory. More importantly, it provides recommendations and identifies the risks associated with not carrying out the recommendations. These recommendations range from informative to critical, helping you to prioritize your actions accordingly.
How Can I Start Monitoring Fast?
While VIOS increases flexibility and helps keep hardware costs low, it can feel unfamiliar to IBM i users. This 30-minute recorded webinar introduces easy monitoring templates that can help you maximize the benefits without AIX expertise.