Gateway Unresponsive

General Discussion
1

I won’t have all the details until Monday (3/30), but wanted to go on and make this post in case anyone has seen this before:

Igntion 7.5.xx (latest version)
Windows 2008 Server

Attempting to connect to 40+ Allen-Bradley PLCs, mostly on DH+ via a ControlLogix gateway.

After about 25 connections are made, the Igntion Gateway becomes unresponsive.

Any suggestions are greatly appreciated. I will post more information on Monday.

2

More information regarding post:

Wrapper logs attached. Start at the end, many of the issues early in the files have been addressed and fixed.

We are migrating a license for 7.5.x from an older machine to a newer machine.

After installing Ignition on the newer machine we could not launch a designer, found that Java 8 was installed on the server. Uninstalled Java 8, uninstalled Ignition, installed Java 7, installed Ignition. All appears to be working fine.

Customer started adding devices, most of which are AB PLC5 and SLC attached to a ControlLogix gateway via DH+. After adding approximately 40 devices, the gateway became unresponsive (3/28 15:17).

We resolved the database issues and “Disabled Processor Browse” for all devices.

Now we start getting connection issues at about 22 devices.

Let me know if additional information is required.

Thanks for any assistance.
wrapper.zip (685 KB)

3

You might try increasing the memory allocated to Ignition. You can find this in the ignition.conf file int he installation directory. If you’ve got a 64-bit machine with 64-bit Java installed start at 2GB and see if that helps.

4

Thanks Kevin.

Do you think increasing the memory will fix the device connection issue?

5

[quote=“mcgheeiv”]Thanks Kevin.

Do you think increasing the memory will fix the device connection issue?[/quote]

It might. You’ve got a lot of “clock drift detected” warnings happening, which can lead to things like devices timing out because basically what’s happening is the system is freezing up for multiple seconds at a time. While not the only cause, not having enough memory allocated and causing the JVM garbage collector to be constantly overworked is probably the most common cause.

6

This will fix it. We hit a similar issue recently.

7

FWIW, I’ve had spectacularly good results from the G1 garbage collection in java7 and 8. Handling multiple Gig allocated in a system with tens of thousands of tags, it can hold stop-the-world pauses to under 100ms, and typically under 50ms.