Version

Argent Advanced Technology all versions

Date

Wednesday, 20 Nov 2013

Summary

Slow file shares on the Main Argent Engine can cancel tasks when Argent Advanced Technology Trusted Agents are used, as well as slow-loading Argent SuperMaps and Enterprise Application Views in Argent Commander

Customers using Non-Stop Motors may also notice synchronization issues

Technical Background

Slow file shares can be caused by TCP port exhaustion – this is when the number of TCP ports available are exhausted either due to Windows licensing restrictions, TCP settings for the maximum number of ports, Group Policy, network card restrictions or hardware issues

When the number of TCP ports exceeds a certain amount (varies on different machines and editions of Windows), access to file shares can take anywhere from 10 seconds to 10 minutes

File shares are critical in the operation of Trusted Agents for fetching work-orders, and required for Argent Commander to access the SuperMaps and Enterprise Application View XML files to render the images in the browser

When TCP ports are exhausted, a ‘queue‘ builds-up – the file share request suddenly needs to wait until there are available sockets to open, which severely delays the ability to connect to the file share.

On loaded systems, a file share request could take up to 10 minutes as the queue is populated by other pending socket requests

Customers can see which TCP sockets have been established by typing:

netstat -n > RESULTS.TXT

Customers can open the generated RESULTS.TXT file and view the total line count in Notepad’s status bar to get a rough idea of how many TCP sockets are in use

Resolution

TCP port exhaustion can occur for multiple reasons – possible changes in Argent architecture (e.g. changing from Trusted Agents to Daughter Engines) are one approach

Daughter Engines use only a single TCP connection to receive and send data while Trusted Agents require multiple connections for each task or ‘work order‘ that it is processing concurrently

However, if the TCP connections are caused by non-Argent-related applications (e.g. Enterprise anti-virus, backup programs, patch management systems, etc.) – then Argent can assist with identifying precisely which application is generating the highest TCP load