View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0004586||SymmetricDS||Bug||public||2020-10-15 01:50||2020-10-15 19:27|
|Summary||0004586: When sym_table_reload_status has more unprocessed rows than initial.load.extract.thread.per.server.count, loads don't queue|
|Description||When sym_table_reload_status has more unprocessed rows than initial.load.extract.thread.per.server.count, loads don't queue from sym_table_reload_request.|
[Edit] Chris said that setting start.initial.load.extract.job=false is needed to reproduce this. We'll give that a try.
|Tags||No tags attached.|
||This is working as designed. There is a warning in the log and on the screen when you have more loads requested than threads. It is more efficient to process a load to completion than to task switch across multiple loads. Customers were complaining that loads were too slow, but it was the multi-tasking that made it appear slow, so it was changed to dedicate threads to loads. If you want more initial loads in parallel, then increase the thread count.|