Hi,
I am currently having an issue getting data from site (running ignition edge) to a central ignition server. Specifically I am seeing a number of problems:
- The disk is showing > 100% storage, and keeps going up. At some point, it somehow drops down to 100% although i have never been watching when this happens so i don’t know if there was anything special about when it occurs. I have seen it creep up to about 130%. It never goes below 100%.
Admittedly, I am sampling a couple of hundred tags with a 1 second scan class, with a poor cellular connection but i have tried other sample rates, which, although have the effect of moving the disk storage to idle, and thus implying that the data stream is not being held up, the disk cache never empties itself, so i don’t know what the buffer is full of or why it is not being forwarded now that it has the opportunity to. - Altering the historian scan class that i am using for all historical tags does seem to have some impact on the throughput to the database, but has no impact on the rate at which the disk cache is being filled, which is still between 10-70 records per second (it fluctuates, but i have no idea based on what). Because the batched pushes to the central gateway, and the storage into the disk cache are not synchronized, i assume all data is written to disk to be batch processed later. Therefore the throughput to the DB should be subtracted from the throughput to the disk, to give the net throughput, but instead, the disk cache keeps getting larger.
- This problem is repeated on about 6 different sites with the same setup, all behaving the same, and all have a slow connection.
Admittedly, I am struggling to understand how to interpret the numbers as my understanding of how the store and forward works on ignition edge doesn’t marry up with what i am observing, so i may need some education on the matter too, if what I have said doesn’t actually make sense.
I can play with scan classes etc and reduce the required throughput, but i really need to understand why it is doing what it is doing in the first place, particularly, if scan classes are having no effect on the volume of the disk cache being used.