samueldlightfoot commented on PR #4606: URL: https://github.com/apache/cassandra/pull/4606#issuecomment-3932984509
> TY, looks pretty good. Do we have testing of what happens when the configured mode is direct IO but direct IO isn't supported? What I am wondering is if there are any crossed wires where the access mode on the file doesn't match the actual way it was opened. > > I don't think this is the case just because it's the dfile that is used. > > The other important thing is to look at the tests that were added in the cursor based compaction patch and then parameterize those with direct IO so it is exercised. > > Really the tests you parameterized should maybe also be parameterized on cursor based compaction as well? I am assuming those tests are relatively fast executing so it should be fine? Thanks. The newly added _SSTableReaderDataReaderTest_ covers the fallback cases and asserts the same channel (fd) is used in such. I can't think of a way the access mode on the FileHandle could be disjoint from the way the file was opened. I've parameterised the main compaction tests across cursor/iterator and standard/direct (they were all quick to run). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
