Best Practices for Deploying the Sun StorADE Utility
- Sun StorADE 2.2 Features
- Installing StorADE
- Related Resources
- About the Authors
This article discusses best practices for deploying the Sun Storage Automated Diagnostic Environment (StorADE) utility to monitor Sun supported storage devices in the data center. The purpose of this to provide StorADE installation and configuration best practice recommendations that result in the optimal use of the StorADE utility based on the storage topology of a given environment
The first part of this article provides a high-level overview of the StorADE utility, and describes which environments are best served by the utility. This information is relevant for IT architects, or anyone looking for ways to improve the reliability, availability, and serviceability (RAS) of their Sun storage solutions.
The second part of this article provides step-by-step installation and configuration information that can be used by an administrator, or anyone tasked with deploying the StorADE utility. The information provides helpful suggestions for making the most of your StorADE deployment, whether the environment contains complex storage area networks (SANs), or straightforward direct-connect devices.
The following topics are covered in this article:
"Sun StorADE 2.2 Features"
"Installing StorADE"
Sun StorADE 2.2 Features
StorADE is a utility provided by Sun that enables centralized monitoring of most Sun storage product offerings. For a complete listing of supported products, see TABLE 1.
StorADE uses well-established Solaris™ Operating System (OS) facilities, such as crond and inetd, for operation. StorADE use and administration is performed using standard browsers such as Netscape Navigator™, Mozilla, or Microsoft Internet Explorer. Refer to the StorADE release notes for supported versions.
StorADE 2.2 is available at no cost as a download from: http://www.sun.com/downloads. For details on downloading this utility, see "To Download the StorADE Package" on page 14.
The Sun StorADE utility has the following features:
-
Enables centralized monitoring for all supported devices and hosts through a browser-based user interface (BUI). You view the devices by topology or by list (FIGURE 8).
Monitors storage devices, providing a health status at the storage subsystem, device, and field replaceable unit (FRU) levels.
Delivers notification of key events in the following ways (based on configuration):
Local notification Events can be communicated to a local administrator (through email), or to local management software (such as Sun™ Management Center, or through scripts to third-party applications).
Remote notification Events are communicated to Sun using Sunsm Remote Services Net Connect service (SRS Net Connect) for improved serviceability.
Maintains logs and provides on-demand reports for storage device events.
Provides storage diagnostics that enable a point-and-click approach to fault isolation.
Offers expert advice and troubleshooting for actionable events.
Checks the software, hardware, and firmware revisions of all monitored storage devices and hosts.
Provides detailed storage device FRU reports (FIGURE 1) that describe the device type, vendor, model, and revision number information.
Offers a secure-mode configuration for environments where security is a concern.
FIGURE 1 shows the StorADE FRU and model summary page.
FIGURE 1 StorADE FRU and Model Summary
Supported Storage Products
Once installed, StorADE monitors the local and remote storage products listed in TABLE 1.
NOTE
The product list in TABLE 1 is up to date as of the time of publication. Refer to the StorADE release notes for updates.
TABLE 1 Supported Products
Sun StorEdge™ products |
Sun StorEdge A5000 array (14 slot) |
|
Sun StorEdge A5200 array (22 slot) |
|
Sun StorEdge A3500-FC array |
|
Sun StorEdge T3 array (F/W 1.17 or greater) |
|
Sun StorEdge T3+ array (F/W 2.0 or greater) |
|
Sun StorEdge 6120 array |
|
Sun StorEdge FC Tape T9840 |
|
Sun StorEdge FC Tape T9840B |
|
Sun V880 Server |
|
Sun StorEdge D2 Array |
|
Sun Internal Fibre Channel disks (E3500) |
|
Sun StorEdge 3510 FC array |
|
Sun StorEdge 3900/6900 series |
|
Sun StorEdge 6320 system |
|
Sun StorEdge Solution racks |
SAN-type products |
Sun StorEdge Network FC Switch-8 and Switch-16. |
|
Sun StorEdge Network 2-Gbit FC Switch-8, Switch-16 and Switch-64. |
|
Brocade Silkworm 2400/2800 FC Switch |
|
Brocade Silkworm 3200/3800/12000 FC Switch |
|
Brocade 3900 32-Port FC Switch |
|
McData ED6064 FC Switch |
|
McData ES3232 FC Switch |
|
McData 4500 2GB 24-Port FC Switch |
|
McData 6140 FC Switch |
|
Inrange FC 9000 Director FC Switch |
Host bus adapters |
Sun StorEdge PCI FC-100 Host Adapter |
|
Sun StorEdge SBus FC-100 Host Adapter |
|
Sun StorEdge PCI Dual Fibre Channel Host Adapter |
|
Sun StorEdge 2 Gb FC PCI Single-Channel Network Adapter |
|
Sun StorEdge 2 Gb FC PCI Dual-Channel Network Adapter |
|
Sun StorEdge 2 Gb FC cPCI Dual-Channel Network Adapter |
|
JNI FC Host Adapter |
TABLE 2
How StorADE Works
The StorAdE software is installed on one or more servers in a data center. During the installation you are prompted to configure the server as a StorADE master agent or slave agent. The following points and FIGURE 2 describe the roles of each.
Master Agent serves the following functions:
Aggregates all monitoring data collected from master and slave monitoring activities.
Monitors devices out of band (monitors the device using the device Ethernet interface).
Monitors devices in band (monitors the device using the device Fibre Channel or SCSI data path.)
Monitors various Solaris OE log files for storage-related events.
Stores StorADE events in StorADE log files.
Sends notification of key events to local personnel using simple mail transfer protocol (SMTP).
Optionally, sends notification of key events to Sun using SRS Net Connect. For details on SRS Net Connect services, go to http://www.sun.com/srs.
Provides the centralized BUI for StorADE administration and for checking monitoring data.
Can be used as a management station (performs no monitoring; used only to aggregate data, send notifications, and provide the BUI).
Slave Agent serves as another StorADE resource with the following functions:
Provides the ability to scale the StorADE functionality according to needs.
Monitors devices out of band (monitors the device using the device Ethernet interface).
Monitors devices in band (monitors the device using the device Fibre Channel or SCSI data path.)
Monitors its own Solaris OE log files for storage-related events.
Stores StorADE events in StorADE log files.
Forwards actionable events and alerts from all monitoring activities to the master agent.
Can be installed on a log host, and used only as a central repository for log files.
Installing the StorADE package on a server adds a cron entry to the server and a StorADE-specific http service. The cron wakes up the StorADE agent periodically (the default is five minutes) to probe devices and monitor log files.
The StorADE user interface is accessed through a browser. Pointing a browser at a slave host port 7654 (or port 7443 for secure configurations) redirects it to the master server automatically.
The StorADE Service Advisor feature provides content that includes a summary of the event, its probable-cause, as well as a recommended corrective action.
FIGURE 2 StorADE Configuration Components
Deploying StorADE
There are a variety of ways to deploy the StorADE utility in your environment. Your best practice StorADE deployment depends on the type of storage you plan to monitor and the topology of your enterprise.
This article outlines four sample monitoring solutions for entry-level and mid-range servers. These solutions correspond with the overall environment (topology) rather than with particular applications running in it. You can use any of these deployment solutions as is, or tailor them to fit your exact needs. Examples of customization opportunities include adding additional slaves or data hosts to distribute load and scale the StorADE utility.
Review the four sample monitoring solutions to identify which solution meets your needs, and use the configuration to plan your StorADE deployment. Large complex environments might benefit by using a combination of the given solutions.
Solution A Single Agent with In-Band and Out-of-Band Devices
The configuration shown in FIGURE 3 is made up of a single master agent installed on one host that is used to monitor everythingin-band devices, out-of-band devices, and log files. The master provides the console and is responsible for sending all notifications. The host must have access to all log files, all out-of-band devices, and connectivity for notifications (such as email and SRS Net Connect).
FIGURE 3 StorADE Deployment for Solution A
TABLE 3 Pros and Cons for Solution A
StorADE Configuration |
1 Master |
Pros |
This configuration is the simplest and easiest to implement. |
Cons |
StorADE has a single point of failure. If the master host goes down, no monitoring or notifications take place. All monitoring is performed by one host. Depending on the number of devices monitored, this configuration might reach saturation, or become negatively affected by a bottleneck (in the host's network interface, for example). |
If your site includes multiple hosts and only one agent is installed, the topology will be incomplete. Consider Solution B if a complete topology is important. If array log files are stored on a separate host, consider Solution C or D.
Solution B Master and Slave Configuration With In-band and Out-of-band Devices
The configuration shown in FIGURE 4 uses one master host that monitors in-band and out-of-band devices, provides the console, and sends notifications. The slave host is used to monitor in-band host bus adapters (HBAs). Log files are managed by the master and the slave hosts. This solution provides a more complete SAN topology because there is an agent on multiple hosts. This solution is well adapted to sites that have multiple hosts with multiple log files.
FIGURE 4 StorADE Deployment for Solution B
TABLE 4 Pros and Cons for Solution B
StorADE Configuration |
1 Master 1 or more Slaves |
Pros |
This configuration provides some redundancy. If one of the StorADE hosts is down, the others continue to monitor devices. The slave host provides log monitoring for out-of-band devices and in-band HBA monitoring. |
Cons |
StorADE is more complex to install than in Solution A. |
Solution C Master and Slave With In-Band and Out-of-band Devices and a Loghost
The configuration shown in FIGURE 5 uses a master host to monitor in-band devices, out-of-band devices, and slaves. A slave is added to monitor HBAs and log files, with another slave installed (as a data host) on a log host server to capture log messages from the Sun StorEdge T3 arrays. For sites where a separate host is used to store log files, this solution can be used to relay the log information to the master agent. Otherwise, Solution C is the same as Solution B.
FIGURE 5 StorADE Deployment for Solution C
TABLE 5 Pros and Cons for Solution C
StorADE Configuration |
1 Master (in-band and out-of-band access) 1 or more Slaves (for in-band coverage only) 1 Slave (for log hosting only) |
Pros |
You get the same advantages as Solution B, with the addition of a data host that provides log monitoring of out-of-band devices. |
Cons |
StorADE is more complex to install than in Solution A or B. |
Solution D Management Station Configuration
The configuration shown in FIGURE 6 uses the master host to receive monitoring information for out-of-band devices from slaves. The master provides the console and sends all notifications. Slaves are needed to monitor HBAs and log files. Multiple slaves are used to scale the monitoring activities. One slave is installed on a log host for log monitoring. When a large number of slaves are present, there is often a need for a separate master to aggregate the slaves and reach the remote monitoring notifiers. When the master's only functions are aggregating the slaves, providing the BUI, and sending notifications, it is often referred to as a management station. This solution is necessary if none of the slaves have access to the remote notification providers (such as SRS Net Connect). This situation can be common on large sites with multiple subnets.
FIGURE 6 StorADE Deployment for Solution D
TABLE 6 Pros and Cons for Solution D
StorADE Configuration |
1 Master (out-of-band log host) 1 or more Slaves 1 Slave (for the data host) |
Pros |
You get the benefits of a management station where you have a central point for the BUI, aggregation of slave monitoring, and logging of out-of-band devices. This configuration provides the best monitoring topology of all four solutions. |
Cons |
This is the most complex solution to install. |
Enabling Security
No matter which deployment solution you use, if security is a concern, you should install all agents with security enabled. StorADE is installed with security turned on when you run the ras_install installation script and answer yes to the security question.
The StorADE secure installation uses the Secure Sockets Layer (SSL) protocol for transmission of information between the master and slaves, and between the master and the StorADE BUI. The highest grade encryption (RC4 with 128-bit secret key) is used. The StorADE package includes a default certificate that expires in 2008. This certificate is located in /opt/SUNWstade/System/certificate.pem.
When secure mode is used, the default URL (port 7653) is used when a redirect occurs. The non-secure URL is http://hostname:7654. Site-specific certificates can be created with the openssl utilities (part of the public domain OpenSSL product).
StorADE also supports user-level security. StorADE user accounts are created by the root (ras) login. Each user is assigned user privileges based on one of the following roles:
guest can read StorADE data
admin can read StorADE data and make StorADE configuration modifications
expert can run StorADE reports and diagnostics
test can run StorADE diagnostics
ras has all the privileges of all the roles (the root equivalent for the StorADE utility)
Each login account allows the user to log in to StorADE with their own login and password, and results in the user having a restricted set of functions available in the BUI.