C H A P T E R 6 |
Disk Control and Monitor Utility (DCMU) for RHEL |
This chapter describes how to use the Disk Control and Monitor Utility (DCMU) on a Red Hat Enterprise Linux U4 (RHEL4 U4 or 4.5) 64-bit operating system. It includes the following sections:
The Disk Control and Monitor Utility (DCMU) controls and monitors all 48-disk drives on the Sun Fire X4500 server and provides the following features:
The Disk Control and Monitor Utility (DCMU) consist of three components. Each component updates the FRU, SDR (Sensor Data Record), SEL (System Event Log) and service processor logs:
To use Disk Control and Monitor Utility (DCMU), you must install the application. To install the application, you must perform the following steps:
The installation of DCMU consists of one step because the package is in rpm format. The DCMU package comes with two rpm files. One is the source rpm and other is the binary rpm:
To Install DCMU |
# rpm -ivh dcmu-1.3-5.x86_64.rpm
The following files are installed as components of the DCMU installation:
The initial installation of the DCMU components prepares the system for running the DCMU utilities described in this chapter. However, since the DCMU utilities also require that the IPMI service is running, you have two options before you can start using the DCMU utilities: Manually start the IPMI service, or reboot the server (which automatically starts faultmond and IPMI).
If rebooting the server after the initial DCMU installation is not possible, and you wish to run DCMU utilities, you must manually start the IPMI service by entering the following command:
Note - After the initial installation of DCMU, rebooting the server starts both IPMI and faultmond. |
If you need to uninstall DCMU, perform the following procedure.
To Uninstall DCMU |
The cfgdisk command is a command-line utility and that queries and provides status of all 48-disk drives located in the Sun Fire X4500 server. cfgdisk also allows you to connect and disconnect disk drives from the OS and also allows you to monitor disks connected to the server.
Use the cfgdisk command to connect, disconnect, and determine disk drive status by using the parameters shown in TABLE 6-1. The following options are supported for the functions shown:
This section contains examples of common cfgdisk commands. For more information and options, refer to the cfgdisk man page.
The following command displays a map of all disk drives:
Here is an example of cfgdisk command output listing physical slot number, logical name, and status information:
Use the cfgdisk command to disconnect a disk before performing the hot plug event of physically removing it. The following command shows an example of how to use cfgdisk to disconnect a disk drive.
# cfgdisk -o disconnect -d sata5/1
The command returns the following prompts; enter Y at both to disconnect the disk:
Are you sure (y/n)? y Are you sure sata5/1 device is not in use(y/n)? y Device sata5/1 has been successfully disconnected. |
After performing the hot plug event of physically adding a disk into the system, use the cfgdisk command to connect it. The following command shows an example of how to use cfgdisk to connect a disk drive.
# cfgdisk -o connect -d sata5/1
The command returns the following:
Command has been issued to connect sata5/1 device, it may take a few seconds to connect sata5/1, check status by re-running cfgdisk command. |
The following command show how to use the cfgdisk command to display help information:
Faultmond is a component of the Disk Control and Monitor Utility (DCMU). Faultmond is a daemon which is started at boot time. It scans all disk at polling intervals, and then reports FRU, SDR, and SEL information to the service processor.
faultmond uses the parameters shown in TABLE 6-2. The following options are supported for the functions shown
This section contains examples of common faultmond commands. For more information, refer to the faultmond man page.
The following command shows the use of faultmond.
The command returns the following:
To start faultmond, enter the following command:
To stop faultmond, enter the following command:
To set the polling interval with faultmond, do the following:
1. Stop faultmond from the command line.
2. Set the polling interval. For example, to set the polling interval to be 1 minute, you would enter:
3. Check the polling interval.
The output looks like the following:
# ps -ef |grep faultmond root 15357 1 5 15:49 ? 00:00:00 faultmond -t 1 root 15364 15307 0 15:50 pts/4 00:00:00 grep faultmond |
hotplugmon is not a command line utility. It monitors hotplug events and then reports them to service processor.
Note - hotplugmon is only activated with faultmond from the command-line or during boot time. To stop or start faultmond and hotplugmon manually, you should use the faultmond service commands. |
As described above, DCMU monitors hotplug events and pending drive failures, controlled connect/disconnect events and logs these events in syslog and, more importantly, in the service processor logs (SDR, FRU, SEL). You may access these logs individually for specific information to aid in the administration or troubleshooting of the disk array. This section describes how to view individual log file information from the command line.
The following commands show how view the SDR log file, either at the server:
# ipmitool -I lan -H SP-IP -U root -P SP-password sdr elist
Where SP-IP represents the IP address of the service processor and SP-password represents the password for the service processor.
The following commands show how view the FRU log file, either at the server:
# ipmitool -I lan -H SP-IP -U root -P SP-password fru
Where SP-IP represents the IP address of the service processor and SP-password represents the password for the service processor.
The following commands show how view the SEL log file, either at the server:
# ipmitool -I lan -H SP-IP -U root -P SP-password sel elist
Where SP-IP represents the IP address of the service processor and SP-password represents the password for the service processor.
All events and error information from DCMU are logged in syslog (default: /var/log/messages). These include hard drive hotplug events, drive disconnect and connect events, and drive fault polling events.
Copyright © 2009 Sun Microsystems, Inc. All rights reserved.