Starting IP/MPLSView Servers
Before using IP/MPLSView, you need to start up the IP/MPLSView server and connect to it from an IP/MPLSView client. The IP/MPLSView server software runs on a Linux machine; the client software can run either on a Linux machine or PC with Microsoft Windows. This chapter explains how to start up the server and connect to it with the client.
Starting Up the IP/MPLSView Server
Before starting the IP/MPLSView server or client, install them. For more information, see Installing the IP/MPLSView Server, Client, Traffic Data Collector, and Rsync Package.
Table 1 provides a summary of the commands used to start, stop, and determine the status of IP/MPLSView servers. Do not use root to startup the IP/MPLSView servers because this might lead to ownership and process conflicts.
Table 1: Starting, Stopping, and Verifying Status Commands
Command |
Explanation |
---|---|
status_mplsview |
Use this command to check the status of the IP/MPLSView servers. This command displays IP/MPLSView process information for CPU/Memory usage, as well as any warnings or errors regarding the status of IP/MPLSView-related processes. The process information at the beginning of the status output
can be used to fine-tune the server memory usage settings, which can
be configured using All errors displayed are fatal. They indicate that the server or the IP/MPLSView data collection driver is not running. Any warnings that are displayed at the end should warrant attention. |
startup_mplsview port-number |
Use this command to start all the servers. The startup script either confirms the startup process if the IP/MPLSView servers started successfully or provides error message(s) if the start server process fails. Do not use root to startup the IP/MPLSView servers. |
stop_mplsview |
Use this command to stop all the servers. Note: This command gives the IP/MPLSView clients a one minute notice in which to save unfinished work. |
nodejs_server.sh start |
Use this command to start the IP/MPLSView servers if you are using the RESTful API through the Web-based UI. The Node.js process must be up for the Web UI to be accessible. |
nodejs_server.sh stop |
Use this command to stop the IP/MPLSView servers if you are using the RESTful API through the Web-based UI. The Node.js process must be up for the Web UI to be accessible. |
The ./nodejs_server.sh start and ./nodejs_server.sh stop commands start and stop the IP/MPLSView servers if you are using the Representation State Transfer (RESTful) API included in IP/MPLSView. The RESTful API enables northbound systems to receive various sets of data by means of standard HTTP requests.
The “Node” product (Node.js) receives and processes RESTful API HTTP requests on the IP/MPLSView application server. It uses a port other than the standard 8091 port used by the application. The default port for this interface is 3000.
- Before starting IP/MPLSView, contact Juniper support to
obtainan activation license file. You need the host ID of your server.
To determine your host ID:
hostid
Change the license filename to npatpw and move it to the /u/wandl/db/sys folder, or the
$WANDL_HOME/db/sys
folder.You can also upload the license file from the Web Interface to the application server and view it from the Web Interface. This feature is available from the Web menu by selecting Admin > Administration > Upload License and Show License.
- Log into the server as the IP/MPLSView admin user; the
wandl user ID is recommended when applicable. After you log in, change
directory using the following command:
$ cd /u/wandl/bin
- Verify that the IP/MPLSView-related servers are running
using the following command.
./status_mplsview
- If the servers are not running, start them using the following
command:
./startup_mplsview
The IP/MPLSView server and Tomcat server (a web application server) should be listed in the servers that are running. The default port number for the IP/MPLSView server is 7000 and is used for communication between the server and client. When you startup IP/MPLSView, the IP/MPLSView data collection driver (wDriver) also takes port 7001.
- If the default port is in use, you can start the server
with another port number by using the command:
$ ./startup_mplsview port_number
In this example, substitute port_number with a valid unused port number. The IP/MPLSView data collection driver port number is then set to port_number+1.
- You can optionally set up the PATH variable to access
the commands located in
/u/wandl/bin
and/u/wandl/client
folders from other directories without specifying the full path. The following variable can be placed in the$HOME/.profile
file.PATH=/u/wandl/bin:/u/wandl/client:$PATH export PATH
-
After setting the PATH variables, you can use the settings in the
.profile
file, using the following:# . ./.profile
![]() |
Caution: If you choose to setup the PATH in the .profile file, remember to change the PATH when reinstalling IP/MPLSView in a different folder. Otherwise, an older version might be executed from the directory in the path instead of the version from your current directory. You can check which version you are running by using the which command. |
Starting the IP/MPLSView Traffic Data Collector
The Traffic Data Collector should be started on the server if you want to collect traffic data using the IP/MPLSView client.
To launch the Traffic Data Collector application:
- Switch to the directory where the Traffic Data Collector
is installed. For example,
/u/wandl/dcollect
, and then execute the dc.sh script.$ cd /u/wandl/dcollect
$ ./dc.sh start instance#
In this example, instance# can be any positive integer representing the instance number of the collector that is being started. There can be more than one instance of the Traffic Data Collector running at once on the same server or on other servers. You should see some messages similar to the following:
Trying to start using pid=8608 Data Collector (pid=8608) Started.
- If a Traffic Data Collector is on a different machine
than the main IP/MPLSView application server, specify the IP address
of the application server using the -h option:
$ ./dc.sh start instance# -h host_ip_address
Multiple instance# can also be started in sequence using a comma or a number range.
The following example starts instance 2, 3, 5, and 7.
$ ./dc.sh start 2,3,5,7
The following example starts instance 0, 1, 2, 3, 4, and 5 inclusive.
$ ./dc.sh start 0-5
- Run the following command to see the status of the collector:
$ ./dc.sh status
Found collector instance wandl_0 with pid=8608 (running) Found collector instance wandl_1 with pid=8628 (running)
In this example, wandl_# represents the instance identifier of that Traffic Data Collector and wandl represents the user ID that started the Traffic Data Collector.
After verifying the status of the collector, you can specify what to collect using the Traffic Collection Manager in an IP/MPLSView client. The client can be on a different machine.
- To stop a specific Traffic Data Collector instance, use
the following:
./dc.sh stop instance#
- To stop all Traffic Data Collector instances, use the
following:
./dc.sh stop all
- If an instance is having difficulty stopping, you might
force a shut down using the -f flag.
./dc.sh -f stop instance#
Starting the IP/MPLSView Trap Daemon from Linux
When running the startup_mplsview or stop_mplsview scripts, you might be asked to stop or start the SNMP trap server. The SNMP trap server is used for the online Fault Management module.
The SNMP trap daemon can also be manually started on the server if you wish to view traps using the IP/MPLSView client, such as through the event browser.
To start the SNMP trap daemon:
- To launch the trap daemon, first change directory to the
bin directory where IP/MPLSView is installed (for example,
/u/wandl/bin
) and then use the .snmptrap start command.$ cd /u/wandl/bin
$ ./.snmptrap start
SNMP trap daemon started on port 162
- Before you can stop the SNMP trap daemon, first stop the
application monitor if it is enabled.
$ ./.appmonitor stop
If the application monitor is not stopped first, it might detect that the SNMP trap server is down, and automatically restart it.
- Stop the SNMP trap daemon.
./.snmptrap stop
Manually Starting the Event Server
Note that the event server is automatically started up when you use the /u/wandl/bin/startup_mplsview command. There might be circumstances when you need to manually stop and restart the event server.
- To manually stop and restart event server:
$ cd /u/wandl/bin
$ ./.appmonitor stop
$ ./.eventserver stop
$ ./.eventserver start /u/wandl/db/config/eventserver.xml
$ ./.appmonitor start
By default, the event server only displays application events. In order to also display SNMP trap events, start the trap daemon as explained in Starting the IP/MPLSView Trap Daemon from Linux.
Manually Starting the Distributed Database in Linux
In most cases, the database is installed with the IP/MPLSView
server. If so, starting up the IP/MPLSView server also starts up the
database. However, if the database is installed on a different machine
from IP/MPLSView, change directory to the bin directory of where the
distributed database is installed. For example, /u/wandl/bin
.
- To start up the database:
$ startup_mplsview
- To shut down the database:
$ stop_mplsview
Restarting the Task Server
Under special circumstances, you might want to stop and restart the Task Server only, without taking down other processes. The application monitor should be stopped beforehand, so that it does not automatically restart the Task Manager when it detects that it is down.
- To stop the Task Server process, use the following commands:
$ cd /u/wandl/bin
$ ./.appmonitor stop
$ ./.tmng stop
- To restart the Task Server process, use the following
commands:
$ cd /u/wandl/bin
$ ./.tmng
$ ./.appmonitor start