800xA 5.1: Property Transfer: "write failed for x properties" and " Transaction overrun for property ..."
In 800xA 5.1Fp4 rev E, I have an analog value from an AC450 (MB300). I'm using Property Transfer to write that value to a third-party OPC server (KepWare) (Maximum rate set to 10 000 milliseconds). I also use Property Transfer to transfer a few other values to and from General Property aspects. I get quite a lot of these events: "Write failed for 1 properties" and "Transaction overrun for property ... overrun in x seconds". You would think the Property Transfer service is overloaded, but given the low amount of signals I'm transferring, that really shouldn't be the case. I've tried adjusting the properties of the Property Transfer Definition aspect, but nothing seems to help. How do I get rid of these events?
Answers
Never use 10 second subscription times for any Advant Controller .... 3 secs, 9 seconds or more than 30 seconds.
If "quite a lot" is once or twice per day, then ignore it. If "quite a lot" is every few minutes then you need to look at the total number of OPC transfers happening per scan etc.
If "quite a lot" is once or twice per day, then ignore it. If "quite a lot" is every few minutes then you need to look at the total number of OPC transfers happening per scan etc.
The PT service is hardly ever overloaded by its simple job of moving OPC messages between queues.
An overrun occur when a new input is received before write acknowledge from the previous output.
A write failure is what it is; the destination complaining.
An overrun occur when a new input is received before write acknowledge from the previous output.
A write failure is what it is; the destination complaining.
- Make sure most recent ABB 800xA Base Rollup is installed.
- Make sure error messages are received only from the PT provider in Service state. Messages from the Standby are false positives. Cycle the standby PT provider to temporarily work around this bug.
- Use afwapplogviwer (see attached file) to enable the OPC Communicator Common log to level 3 in the running PT provider. Study the expected perpetual sequence of NewData, AsyncWrite and WriteComplete along with the return codes from the write ops (should always be zero, 0). Hint: disable secondary PT, use LogInit to record from very start of primary, save output to file and use a tool like the excellent Notepad++ to filter for itemIDs, codes, etc. From your image it looks like the destination is occasionally returning an error code or is too late to report write is completed. Work these errors/delayed replies with the vendor of the destination. If ABB AC450, performance in write to AC450 is known to be "low", especially if CPU load is high. NEVER write to any other AC450 object type than MB300 DAT (if needed, use PC and MOVE blocks to relay the data to the final destination).
Add new comment