STCLive: Why are all tests failing with "Failed to connect to "ip" "port" " errors but can manually telnet to the IP address and Port noted in the error?

Doc ID: FAQ18365

Printer Friendly Email Article Link

STCLive: Why are all tests failing with "Failed to connect to "ip" "port" " errors but can manually telnet to the IP address and Port noted in the error?

Environment/Versions

Lumos/STCLive-all versions

Answer

If error is happening to any site and you are able to telnet to the IP address and the Port noted in the error, the most likely cause is DeviceControl ran out of file descriptors. To verify do the following:

Log on to the server as user smartsys.
Identify the DeviceControl pid:
cd $METRO_HOME/scripts
statusMetro
Run lsof against the DeviceControl pid and pipe it to word count:
$COTS_PATH/lsof/"version"/lsof -p "pid" | wc -l
where: version is the latest version of lsof.
Pid is the DeviceControl pid.
Check to see how many file descriptors the smartsys user is setup for:
ulimit -a
The value interested in is: nofiles(descriptors)
If the lsof word count is at or close to the ulimit nofiles setting, this problem has occurred.

To fix:

Increase the nofiles limit for the smartsys user (for large systems we recommend minimum value of 4000). The method of setting the value varies based on the type of operating system. Consult your OS admin for steps to increase.
After the OS admin increases the value, log off of the server and back on. Run 'ulimit -a" and verify you see the increase value for the nofiles setting. If not, pass back to the OS admin till it is increased.
Stop and restart all Spirent Applications.

Version 1.0

Publish Date 2018-03-30

Categories
Product : VW Controller PM Classic,VW Controller T&D Classic

Customer Service Center

Spirent KB Article

STCLive: Why are all tests failing with "Failed to connect to "ip" "port" " errors but can manually telnet to the IP address and Port noted in the error?