We're currently using H2 version 199 in embedded mode with default nio file protocol and MVStore storage system. In some installations we encounter DB corruption within minutes after batch operations (a mix of about 30.000 deletes, inserts and updates).
Within 45 seconds of the last insert and roughly 90 seconds before the app detects that the DB is corrupt , we see the following two lines in the log file:
org.mortbay.io.nio.SelectorManager$SelectSet@3180ea95 JVM BUG(s) - injecting delay 3 times
org.mortbay.io.nio.SelectorManager$SelectSet@3180ea95 JVM BUG(s) - recreating selector 3 times, canceled keys 51 times
Is H2 responsible for these log messages? Are they related to the database corruption?
No, H2 is not responsible for those messages. Who is? My 30 sec. Google search shows Jetty as a possible culprit.