AC 800M OPC server fault "Failed to extract file from PPA via FSD"
We found this controller unhealthy (sad) in AC 800M OPC server window on one of the connectivity server in a redundant pair, the second server was okay.
When trying to reconnect to this controller in OPC server we received the following error message:
OPC server log contains the following repeating error messages:
W 2018-06-16 00:57:09.564 Exception when extracting FSD file N4.cdobXml. HRESULT=0x8abb45c5 (ExtractCompilationFile(...)) E 2018-06-16 00:57:09.565 Failed to extract file from PPA via FSD. Name: N4.cdoXml W 2018-06-16 00:57:09.638 Exception when extracting FSD file N4.rrsbXml. HRESULT=0x8abb45c5 (ExtractCompilationFile(...)) E 2018-06-16 00:57:09.638 Failed to extract file from PPA via FSD. Name: N4.rrsXml W 2018-06-16 00:57:36.739 Exception when extracting FSD file N4.ctdbXml. HRESULT=0x8abb45c5 (ExtractCompilationFile(...)) E 2018-06-16 00:57:36.739 Failed to extract file from PPA via FSD. Name: N4.ctdXml W 2018-06-16 00:57:58.091 Exception when extracting FSD file N4.cdobXml. HRESULT=0x8abb45c5 (ExtractCompilationFile(...))
When we tried to restart the faulty connectivity server, we noticed that communication is restored on all operator stations when the faulty server goes down, and lost again when the server is up.
Then we had to disable the OPC DA service provider for the faulty connectivity in the service structure to let DCS operators run the plant.
After a few days during plant shutdown, we enabled the OPC DA service provider back, and everything was okay.
What could cause this failure?
Why we lost communication with the controller at all operator stations? If I understand the Connectivity servers affinity setting correctly (see below), the connection should remain at half of the operator stations.
There can be only one application called 'N4' in controllers connected to the same OPC Server!
I have seen problems like this before, and the problem usually present itself when a connectivity server is restarted.
Please have a proper support case registered with your regional ABB support center.
The OPC server need to be able to exchange data with the Control Builder M stations. This is done via FSD.
Without this data, an OPC server may go ”blind” for one or more controllers.
The OPC DA framework of 800xA will keep an OPC server connected until it flags itself ”bad”. The AC 800M OPC server flags itself out when connection has failed with ALL controllers. Hence, a partial loss where only one OPC server (and its set of clients) is having trouble is possible. Your countermeasure to shut down the OPC DA Connector service provider for the troublesome OPC servers was effective.
Until this problem is corrected, I suggest (rather than stopping an OPC DA service provider and hence also remove redundancy) to try performing a new download to the affected controller (with some change made). FSD will then attempt a new sync.
Stopping AfwFsdNode.exe and erasing the FSD client cache (\OperateITTemp\FsdClient), or even restarting the affected FSD server service provider have helped some in the past.
Ask support if a more recent and corrected version is available (I know corrections have been made in this region in later 5.1 revisions/rollups).