This file describes the Intel® Linux* LAN Adapters PROCfg Application, version 1.4.x.
PROCfg is a reporting and configuration tool for Intel® PRO LAN adapters (10/100 and 1000) and Intel Advanced Network Services (iANS). It works with the Intel e100, e1000, and iANS drivers on Intel 32-bit architectures running Linux.
The tool consists of two parts: the daemon PROCfgd and the PROCfg command-line tool. The daemon should be run on a managed machine. The command-line tool should be used on a managing machine. The managing machine can be either the same managed machine that is used for the daemon, or it can be a remote machine.
PROCfgd and PROCfg require the following minimal versions:
Linux 2.4.x kernels
Any combination of the following Intel drivers:
e100 version 2.1.x
e1000 version 4.4.x
iANS version 2.0.x
![]() |
NOTE: The non-Intel driver eepro100 must not be loaded on the managed machine. |
libstdc++-libc6.1-1.so.2 package - installed by default in Red Hat* 7.2 and later
openssl* 0.9.6 or higher security package installed on the system.
![]() |
NOTE: openssl is installed with Red Hat Linux 7.0 and above. If the package is already installed, make sure the libcrypto.so.1 file is in /usr/lib. If not, see the instructions below for creating libcrypto.so.1. |
If you do not have openssl installed, download the package from http://www.openssl.org and configure openssl with 'config shared' to create the libcrypto.so.0.9.x shared library.
To create libcrypto.so.1 under /usr/lib, you must copy libcrypto.so.0.9.x to /usr/lib and create a link named libcrypto.so.1 to libcrypto.so.0.9.x:
cp libcrypto.so.0.9.x /usr/lib
cd /usr/lib
ln -s libcrypto.so.0.9.x libcrypto.so.1
Ethtool version 1.6 or above, if you have Intel PRO/1000 adapters in your system. This utility is necessary for the 'blink' feature. Ethtool is available at http://prdownloads.sourceforge.net/gkernel/ethtool-1.6.tar.gz?download.
To install Ethtool, refer to the Ethtool INSTALL file instructions. Then, enter the following from the command line:
cd /usr/src/linux-2.4.xx-x/include/linux
mv ethtool.h ethtool.h.orig
cp /root/ethtool-1.6/ethtool-copy.h
mv ethtool-copy.h ethtool.h
If the e100 and/or e1000 drivers are already installed in the system, recompile them with the 'make install' command.
PROCfg includes the following features for Intel adapters:
Viewing:
Bus, Slot, and IRQ numbers for physical adapters
Static driver info: name, version, path, date, size (base and iANS)
Adapter load-time parameters and their values (e100 only)
Dynamic information: link state, speed and duplex (Intel adapters only), and statistical counters
Information for configuring iANS: teaming/VLAN support, adapter capability
Interface information on loaded adapters and teams including virtual interfaces (VLANs) and aliases
The iANS topology (teams, members, VLANs)
Team and member parameter values
Dynamic information regarding the state of a team and the state of its members
Setting:
MAC, inet, mask, and broadcast addresses
MTU size
Interface state
Load-time parameters (e100 only -- when none of the PRO/100 adapters are members of an iANS team or have VLANs)
Team settings: initial config, add members, change teaming mode, change probes settings. Teaming modes include:
Adapter Fault Tolerance ("AFT")
Adaptive Load Balancing ("ALB")
Intel Link Aggregation, Cisco*'s Fast EtherChannel* Technology or static 802.3ad ("FEC", or "FEC/LA/802.3ad: static")
Gigabit equivalent of FEC ("GEC", or "GEC/LA/802.3ad: static")
IEEE 802.3ad: dynamic ("802.3ad")
NONE
VLAN settings: add/delete VLANs for a physical adapter or team (adapters controlled by e100 or e1000 drivers only)
![]() |
NOTE: Once a physical adapter is a team member or has VLANs, any operation issued by ifconfig on the individual network interfaces of such an adapter may cause corruption. As a precaution, Intel recommends using procfg to configure the interfaces and ifconfig only for operations that PROCfg does not support. |
Loading and unloading of drivers
Special Functions:
Blink hardware LEDs to identify a specific adapter
Run diagnostics on adapters
Save and restore network configurations. The configuration is saved as a text file in XML-format. It is possible to edit this file, but restore fails if either the file is in illegal XML format or if it contains invalid configuration data.
![]() |
NOTE: Restoring a configuration destroys the current PROCfg system configuration. This includes unloading Intel drivers. Multi-vendor team (MVT) drivers are not unloaded, as they might control adapters that are not supported by PROCfg and iANS. To avoid MVT adapters using names that are used in the saved configuration file, manually unload the drivers before using the "restore" operation. For additional information, see the Known Issues section. |
Error messages:
User initiated operation failures
Daemon initiated operation failures, usually polling failure
![]() |
NOTE: All error messages are printed to the console. |
Logging:
Configuration changes
If found during periodic polling noted as "external events"
If user change noted as "events generated by session ID X"
Errors that occur only in the daemon
![]() |
NOTE: Events are logged in /var/log/procfgd.log. |
![]() |
NOTE: The following instructions describe installing the PROCfg package through the tar file; although, PROCfg may be installed through a binary RPM* package. To build a binary RPM* package of this application, run 'rpm -tb <filename.tar.gz>'. Replace <filename.tar.gz> with the specific filename of the package. |
Copy the files to a directory of your choice. Put the following files in the same directory:
procfgd-x.x.x.tar.gz (x.x.x designates the version number)
libxerces-c1_6_0.tar.gz (xerces XML-parser library. Needed only if it is not already installed in the system).
Untar the procfgd-x.x.x.tar.gz file:
tar xzvf procfgd-x.x.x.tar.gz
The procfgd-x.x.x directory should appear.
Change directories to the procfgd-x.x.x directory:
cd procfgd-x.x.x
The following files should be in the directory:
INSTALL
INSTALL_BOOT
procfg_init
procfg_init_def
procfg_init_suse
procfg_init_caldera
procfgd
procfgd_adduser
procfgd.1
README
ldistrib.txt
LICENSE.txt
LICENSE.net-snmp.txt
LICENSE.xerces.txt
Enter:
./INSTALL
The INSTALL script untars necessary tar files and copies the following files to the indicated location:
procfgd (the server application) and procfgd_adduser to /usr/sbin
procfgd.1 (the man page) to the default man directory
the licenses, README, and ldistrib.txt files to /usr/share/doc/procfgd-<package_version>
After procfgd is installed, the INSTALL script asks if you want procfgd to automatically run on boot. If you answer 'yes', a script called 'procfgd' is placed in the directory containing the system's boot scripts (usually in /etc/rc.d/init.d/). If you answer 'no' and have later change your mind, you can always enter the 'procfgd-x.x.x' directory and run the command:
./INSTALL_BOOT install
Likewise, if you no longer want procfgd to run on boot, use the command:
./INSTALL_BOOT uninstall
To run the PROCfg server, first logon to the system as root. Enter:
procfgd
![]() |
NOTE: To change the configuration for the PRO LAN adapters or iANS PROCfgd requires a non-default username and password. To configure a user, see the instructions in the next step. |
The daemon runs by default on port 58086. If this port is taken or if PROCfgd is already running on the default port, the following error message is printed on the screen:
init_master_agent: Invalid local port (Address already in use) Server Exiting with code 1
In order to run the server on an alternate port, use the -p option:
procfgd -p <port number>
To add a new read/write-access user to the PROCfg server, first logon as root. Make sure PROCfgd is not running. Enter the following, where the password is at least 8 characters long:
/usr/sbin/procfgd_adduser <username> <password>
Multiple read/write users may be added. The new username(s)/password(s) are added to the /var/.procfg/procfg.conf file. After running the server once, this file is encrypted, but the password is visible while typing it.
![]() |
NOTE: A single read-only user is configured by default on the PROCfg server. To simply view settings no username or password are required, as long as the default username and password (username=procfgd, password=pRoCfGdPaSs) have been left in place. |
Copy the following file to a directory of your choice.
procfg-x.x.x.tar.gz (x.x.x designates the version number)
Untar the procfg-x.x.x.tar.gz file:
tar xzvf procfg-x.x.x.tar.gz
The procfg-x.x.x directory should appear.
Change directories to the procfg-x.x.x directory:
cd procfg-x.x.x
The following files should be in the directory:
INSTALL
procfg
procfg.1
README
ldistrib.txt
LICENSE.txt
LICENSE.net-snmp.txt
Enter:
./INSTALL
The INSTALL script copies the following files to the indicated location:
procfg (the server application) - to /usr/sbin
procfg.1 (the man page) - to the default man directory
the licenses, README, and ldistrib.txt files to /usr/share/doc/procfg-<package_version>
To configure a user in the PROCfg client (procfg), see the Switches section below.
![]() |
NOTE: A single read-only user is configured by default on the PROCfg server. To simply view settings no username or password are required, as long as the default username and password (username=procfgd, password=pRoCfGdPaSs) have been left in place. |
The following parameters are used by entering them on the command line with
the procfg command. When issuing a command, the following syntax must be used:
procfg [<switch(es)>] <command> [<object(s)>] [<option(s)>]
Object(s) can be an interface name (ethx), team name, or list of interfaces or team names (eth0 eth1 ...).
Each switch has a long name as well. See the man page for a more detailed description of the command options.
![]() |
WARNING: The PROCfgd daemon must be running before issuing a command. |
Switch | Description | ||
---|---|---|---|
-p <port> |
If the PROCfgd daemon is not using the default port, you must set the PROCfg application to the same port. If a port number is not specified, the default port (58086) is used. | ||
-h <hostname> |
Hostname or IP address of the managed machine (the one that runs the PROCfgd server). If not specified, the default hostname (localhost) is used. | ||
-U <username> -P <password> |
Set the username and password to access the server application. The username and password
must be pre-configured in the server (see the Installation section). If you do not specify a
username or password, the defaults are used (username: procfgd, password: pRoCfGdPaSs).
|
||
-t <num_seconds> |
This parameter specifies how long, in seconds, the PROCfg application waits for response from PROCfgd. The default is 4 seconds and should only be increased for a highly stressed server. | ||
-f |
"Force mode". Commands do not issue warning to the user. |
To allow a non-default User to write without having to input the username and password on each command, or to save changes to any of these values create a configuration file named procfg.conf. Place it in your home directory under a procfg directory (~/.procfg/procfg.conf). This file may contain the new username, password, port or timeout. The new settings are retained across reboots. An example of the procfg.conf file:
username anyone
password anyonepassword
port 1012
hostname localhost
force
timeout 20
![]() |
NOTE: The configuration file must not contain any white spaces following any of the switches. |
Command | Description | ||
---|---|---|---|
help or -h [<command_in-question> [-a]] |
Displays command usage. Use -a for advanced information. | ||
adapters [<ethx(s) or lspci_adapter_name(s)>] [-a -v -i -h -p[d]] |
Displays general information on adapters in the system including: unique name, system name, link state, speed, duplex, team membership, type (Intel100 for Intel adapters using e100,
Intel1000 for Intel adapters using e1000, vendor names otherwise),
VLANs, interfaces, hardware info, adapter load time parameters.
|
||
adpdiag [<ethx(s)>] [-d <diagnostic_name1> [-p <parameter_value1>]] |
Runs diagnostics on adapters. If no adapter names are given the command displays a list of adapters and their driver, on which a diagnostic can be run. If one adapter name is given without any diagnostic names the command displays a list of diagnostics that can be run on the adapter. If both adapter names and diagnostic names are given the command runs the specified diagnostics on the specified adapters. Specify -p for diagnostics that expect a parameter. | ||
blink ethx [-t <num_seconds>] |
Identifies adapter by blinking hardware LED. Default blink time is 10 seconds.
NOTE: Ethtool blocks all other network configuration while the blink operation is running. |
||
interfaces [<ethx(s)>] [-a -d -s -x] |
Displays general information on all interfaces in the system including:
interface name, inet address, broadcast address, netmask, VLAN ID, MTU size, driver info, and statistics. |
||
teams [<team_name(s)>] [-m -p -v -i] |
Displays information on the teams in the system including: teaming mode, team current primary,
team state, team link state, team speed, team members table, team VLANs table, team
hwaddr, and team probe parameters.
Teams probe parameters are: addressing mode, check time out, send time, max retry count, receive time out, receive back cycles, probe burst size. Team speed is defined differently for the different teaming modes. For AFT and ALB Team speed = primary speed. For FEC and GEC Team speed = sum of speeds of all members. |
||
tree |
Demonstrates the topology of the system. Gives a non-detailed listing of all stand-alone adapters and their VLANs, teams and their VLANs, and members. | ||
vlans |
Displays information on VLANs including: interface name,VLAN name and ID. |
Command | Description |
---|---|
restore [<file_name>] | Restore a network configuration from the host. The configuration is restored from a file in the /etc/procfgd directory. The default file name is saved_conf.procfgd. |
save [<file_name>] | Save the current network configuration on the host. The file is saved in the /etc/procfgd directory. The default file name is saved_conf.procfgd. |
To ensure network configuration is restored:
Ensure that you have installed procfgd with the option to automatically run on boot. For instructions, see the PROCfgd Installation section.
Ensure that procfg is installed on the server where procfgd is installed before implementing the following procedure.
Save the current configuration as the default saved configuration:
procfg -f save
The default file name is /etc/procfgd/saved_conf.procfgd.
![]() |
NOTE: You can run procfg -f save once or multiple times. You can also save multiple configurations to different files using /usr/sbin/procfg save <file_name>. Then, copy the configuration file you want to restore on boot to the default restore file location (/etc/procfgd/saved_conf.procfgd). The file must be saved to /etc/procfgd or any subdirectory under this location. |
Uncomment the following line under /etc/rc.d/init.d/procfgd:
#[ -x /usr/sbin/procfg ] && /usr/sbin/procfg -f restore
During the next reboot, the default configuration will load and restore your network configuration.
![]() |
NOTE: The restore operation overrides network scripts. If you are using DHCP or have other information in the network scripts (such as an IP address), issuing this command will override that information. For example, if you specify an IP address in the script, PROCfg will override that address with the IP address specified in the saved configuration. |
Command | Description | ||
---|---|---|---|
addmem <team_name> <ethx> [-p <priority>] <ethx> [-p <priority>] ... |
Add adapter to team. Default priority is none. | ||
addteam <team_name> eth0 eth1 ... [-M <mode>] [-e/-d] [-a <addrmode>] [-c <num>] [-s <num>] [-m <num>] [-t <num>] [-r <num>] [-b <num>] [-g <aggregation mode>] |
Add team. <team_name> must be no longer than 8 characters, must start with a letter, and must NOT
start with the letters "eth". The default teaming mode is AFT. For other default values, see the man page.
|
||
addvlans <ethx>/<team_name> -i <VLAN_ID(s)> [-n <VLAN_name>] -i <VLAN_ID(s)> [-n <VLAN_name>] ... |
Add VLANs to a team or a physical adapter. Each set of VLAN
IDs can be tagged with a VLAN name. The VLAN ID range is 0 - 4095.
|
||
delmem <team_name> <member_names_list> |
Delete members from team. | ||
delteam <team_name> |
Delete team. | ||
delvlans <ethx>/<team_name> -i <VLAN_ID(s)> |
Delete VLANs from a team or physical adapter. If a VLAN ID appears twice in the list an error occurs. delvlans will not remove VLANs if their interface is up. |
After configuring teaming and VLAN settings, you must save the configuration as the default to make sure it is restored after the next reboot. To ensure network configuration is restored, see the instructions under the Save and Restore commands.
![]() |
NOTE: Enter 'procfg <command>' for a list of the objects that can be configured by the specified command. Enter 'procfg <command> <object>' for a list of parameters that can be set by the specified command and the current values of those parameters. |
Command | Description | ||
---|---|---|---|
adpcfg [<ethx>] [-h <hwaddr>] |
Sets adapter hw addr. | ||
adpsetp [<ethx(s)>] [-p <parameter_name1> = <val1>] [-p <parameter_name2> = <val2>] ... |
Sets load time parameters for adapters that are not in a team or VLAN and that are using the e100 driver.
This operation downs any existing ethx that is using the e100 driver and unloads the driver. Then, the
driver is reloaded with the new parameters, and the previous configuration is saved and restored.
|
||
ifcfg [<ethx(s)>] [-i <addr>] [-n <mask>] [-b <addr>] [-u <pktsize>] [-s <up/down>] |
Sets interfaces (all in list set to same value).
|
||
memcfg <member_name> -p <priority> |
Sets member priority. | ||
teamcfg [<team_name>] [-h <addr>] -[<same_as_addteam_switches>] |
Configures team. |
Command | Description |
---|---|
loaddrv [<driver_name> [-c <insmod_command_parameters>]] |
Loads a driver. If no arguments are given, the operation lists the loadable drivers. If a driver name is given, the operation executes the
insmod command to load the driver. Users may specify additional parameters to pass to the insmod command. See the insmod manpage for a description of the parameters insmod can receive. insmod must include the module name, even if it is identical to the <driver_name> given in the loaddrv command. |
unloaddrv [<driver_name> [-c <rmmod_command_parameters>]] |
Unloads a driver. If no arguments are given, the operation lists the removable drivers. If a driver name is given, the operation executes the rmmod command to unload the driver. Users may specify additional parameters to pass to the rmmod command. See the rmmod manpage for a description of the parameters rmmod can receive. rmmod must include the module name, even if it is identical to the <driver_name> given in the unloaddrv command. |
![]() |
NOTE: For Linux distribution-specific information, see ldistrib.htm. |
An error message, "procfg: Unknown engine id, Unable to connect to host", is seen for every PROCfg command.
Since PROCfg is a client-server application, the loopback interface in the system (lo) must be up, and it must contain an IP address. Configure the loopback interface up by entering:
ifconfig lo <IP_address>
Restoring a configuration destroys the current PROCfg system configuration.
The restore operation might fail in the following cases:
Adapters in the system on which the restore is done are of a different type or have different capabilities than the adapters in the system on which the configuration was saved.
The adapter names in the configuration file are being used by other drivers.
The system is in a state that does not allow loaded drivers to occupy the adapter names that are specified in the configuration file (adapter names are determined when loading the drivers).
In order to overcome these situations either edit the configuration file and change the adapter names to match the system state and the adapter capabilities, or change the system state and/or its adapters in a way which will allow restoring the configuration.
Restoring a configuration does not restore the machine's routing table. Therefore, when issuing a restore operation from a remote machine through a gateway defined on the "restored" machine, communication could stop.
If you have local access to the managed machine, try rebuilding that machine's routing table.
When running PROCfg commands in a script, some commands may fail because of timing issues (some commands take longer than script allows). As a workaround run the script again or insert pauses between the commands in the script.
PROCfg does not warn the user when an unsupported base driver and/or ANS driver is installed on the system. If you are using PROCfg version 1.3.13 or later, ensure that you also have the supported Intel e100, e1000, and/or ANS driver versions as listed in the Prerequisites section of this document.
Last modified on 9/23/02 11:46a Revision 24