Error: Driver output queue to EV manager exceeded (size 10000), starting over writing of Values with same DPID

[shotdunyun]

Hi,

I followed the suggestion of adding smoothing and slowing down poll intervals to devices.
The same issue still happens.

Here's the system description:

SCADA version: WinCC OA 3.12 (no patch)
Server : Stratus ftserver FT4700, running RHEL 6.5
Client: Workstations running Windows 7, single NIC - 10.3.30.218 (client1), 10.3.30.219 (client2)
Server Network configuration (4 NICs):
2 NICs, but bonded to one IP - 10.3.30.3 - ifcfg-bond0 (SCADA network)
2 NICs, but bonded to one IP - 10.5.50.3 - ifcfg-bond1 (Enterprise, SAP network)

Devices on PLC networks : Advantech embedded PCs running RHEL 5 (12 units), PLC (1), Flow Computers (24). (all on 10.3.30.xxx)

Config file:
[general]
pvss_path = "/opt/WinCC_OA/3.12"
proj_path = "/opt/WinCC_OA/Projects/Sxxxxxxx"
proj_version = "3.12"
langs = "en_US.utf8"
pmonPort = 4999
mxProxy = "none"

[valarch]
useHeaderSize = "Linux"


[event]
connectDp = "_Connections"

[driver]

This project was upgraded from PVSS 3.8. Poll rates changed to 400ms for digital points, and 1s for analog points. Smoothing added for analog signals (time dependent, 1s).
It has two WCCOAMod managers, the usual 6 archive managers, one control script manager (only running text parsing scripts collected from flow computers), that is all.
System resources usage is very low. This is a stratus server with hot standby / hot swap capability.
After a few weeks, clients start to lag while the server keeps logging the message queue error. (Error: Driver output queue to EV manager exceeded (size 10000), starting over writing of Values with same DPID).

Please let me know if you have any further ideas / need more information.

Thank you.
Garry

[leoknipp]

Hello,

a detailed explanation for the log-message concerning the driver can also be found in the ETM Portal
https://portal.etm.at/index.php?view=it ... &Itemid=54

I had a quick look at the description for the "Stratus ftServer".
If "hot standby" means that the "virtual machine" is transferred from one server to the other (similar to VMWare vSphere with vMotion) please note that such a solution is officially not supported when using WinCC OA.
Also we have never made tests using a Startus ftServer. WinCC OA supports VMWare Workstation and the ESXi Version 5.1.0 / 5.5.0.


To get detailed information concerning the CPU and memory usage please use the Windows performance monitor. How to configure this software is described in the following FAQ:
https://portal.etm.at/index.php?option= ... &Itemid=54


To get the amount of message (not value changes) processed within the WinCC OA system you can use the WinCC OA message statistic functionality.
A description can be found in the WinCC OA help: System Management --> Diagnostics --> WinCC OA diagnostics.

Best Regards
Leopold Knipp
Senior Support Specialist