Cloud Cruiser became HPE Consumption Analytics on Nov. 1st. You'll still see the old name in places while we update this site.

 

 

Consumption Analytics Documentation

Home > Cloud Cruiser 3 > Setting Up Collection > Universal collectors > Network Traffic Collector > Deployment

Deployment

Initial deployment

To deploy the Network Traffic Collector, perform the following procedure.

  1. Create ccnetcollect.conf with the command ./ccnetcollect -sampleconfig > ccnetcollect.conf

  2. Edit ccnetcollect.conf and specify values for all the required configuration items.

  3. Run the collector using the command ./ccnetcollect

 

If running as a daemon, the configuration file must be in the same directory as the ccnetcollect binary unless otherwise specified by the -f command line option. To avoid confusion when running as a daemon, HPE recommends that you use -f to specify the absolute path and filename of the configuration file.

Post-deployment verification

To verify that deployment was successful, perform the appropriate procedure in this section.

If running as an application (not a daemon)

  1. Verify that the information displayed on screen matches the required configuration. Also verify that there were no errors during initialisation by checking the logfile (or the system log, if no logfile is specified in the configuration file)

  2. Monitor activity by periodically checking the file pc_metrics (in the same directory as the collector), which is updated every 10 seconds. This file contains information about the network traffic and netflow records being processed.

  3. Monitor activity by periodically checking the tail of the file specified by the logfile configuration option, which is updated regularly with timestamped information about runtime events.

  4. Monitor the CCR export directory (by default this is called ccr, in the same directory as the collector) for exported CCR files. Bear in mind that the first CCR file will not be created until the system time reaches the first multiple of the minutes specified by the cc_export_interval configuration item. This multiple is relative to minute 0 of the hour, so if cc_export_interval has a value of 15 then the first export will happen at 15, 30, 45 or 0 minutes past the hour, depending on which is encountered first. 

To stop the collector, press Ctrl+c while it is in the foreground or send it a kill signal if it is in the background, as shown below.

root@PC64:~$ ps aux | grep ccnetcollect
root     44540  1.9  3.3 302656 68400 pts/1    Sl+  14:43   0:00 ./ccnetcollect
root@PC64:~$ kill 44540
root@PC64:~$ !ps
ps aux | grep ccnetcollect
root@PC64:~$

When the collector is stopped, any data held in RAM is lost. Depending on the value of the ccr_export_interval configuration option this may cause up to 59 minutes of usage data to be lost.

If running as a daemon

  1. Verify that the system process list shows 'ccnetcollect' as a running process. This process will not appear in the default 'ps' command output, but should appear in any form of process listing that shows all processes (eg: 'ps -aux' on Ubuntu)
  2. Monitor activity by periodically checking the file pc_metrics (in the directory specified by the daemon_dir configuration option), which is updated every 10 seconds. This file contains information about the network traffic and netflow records being processed.

  3. Monitor activity by checking the tail of the file specified by the logfile configuration option, which is updated regularly with timestamped information about runtime events. If no logfile is defined, then the collector will log to the system log via syslog instead.

  4. Monitor the CCR export directory (by default this is called ccr, in the same directory as the collector) for exported CCR files. Bear in mind that the first CCR file will not be created until the system time reaches the first multiple of the minutes specified by the cc_export_interval configuration item. This multiple is relative to minute 0 of the hour, so if cc_export_interval has a value of 15 then the first export will happen at 15, 30, 45 or 0 minutes past the hour, depending on which is encountered first. 

To stop the collector, send it a kill signal (the default signal SIGTERM sent by the kill command will suffice). The process ID of the daemon is recorded in the logfile when the daemon is started, but can also be obtained by searching the list of processes running on the system in the usual way.

Last modified

Tags

This page has no custom tags.

Classifications

This page has no classifications.
© Copyright 2018 Hewlett Packard Enterprise Development LP