Oracle Rdb V7.1-521 (older version) , waiting for async-prefetch of pages

Causes of excessive async duration reported at a specific time of day???

We RMU unload (read) a specific table every hour (mixed area single table with hash placement index in same storage area). Number of rows in table is low and distributions of data looks random/even. After this unload we then delete the rows (randomly) ready for the next hour. The table is in continuous use propagated using an update trigger from another table in the database. Through the day it is very fast (200,000 pages unloaded/scanned, ~30,000 rows in the unload in a minute or so). I can see it scan the file/table as it unloads. Between 21:00 and 23:59 the