Terasuda,
Both CPU time and CPU Disk Ratio are exceptions that can be defined on a workload in TASM. It would appear that your query exceeded one of those exceptions, which is why you got the error message back.
The CPU Time exception is based on how many CPU seconds have been consumed by the query. This exception is not caused by AMP worker tasks, or being short of AWTs.
The second exception, CPU-to-Disk Ratio, is determined by the number of CPU milliseconds that are being consumed for each I/O. It's intent is to detect queries that are extremely CPU-intensive, such as an unconstrained product join, by detecting how much CPU is consumed for each I/O. It is not caused by AWT exhaustion either. On the face of it, I don't see any direct relationship with getting the error you described and AWT levels.
The most common symptom of being out of AWTs from the query perspective is for a query to take a longer time than expected to complete. It takes longer because one or more of the query steps cannot get an AWT when that step is sent to the AMPS, and so the message representing that step has to wait in the message queue on one or more AMPs. This wait could be subsecond, a few seconds, or a few, or many minutes. This wait in the message queue lengthens the query's elapsed time, but doesn't impact the amount of CPU seconds consumed by the query, or its CPU-to-disk ratio.
↧