Hello there,

    We have used Hornetq (2.2.5) for over a year now without any problems.

    Right now we are upgrading our software and along with that we have changed some Hornetq configurations. This is where the problems started.

    We upgraded the hornet que version to 2.2.14   and changed the  values of the  <page-size-bytes> and <max-size-bytes> (made them both smaller).  (We are going to roll out 2.2.21 in a few weeks now on advise from Clebert  )

   We did not change nor delete journal files , paging files  etc.

After restarting the server we see a lot of : Couldn't locate page transaction xxxxxx, ignoring message on position PagePositionImpl [pageNr=xx, messageNr=xx, recordID=xx

warnings.

Question1:  Since these log entries are classified as warnings I assume they do not affect the integrity of the messages in memory and/or paged messages. Is this assumption correct??? What happens with these messages???  Are they discarded? Wiil they still be consumed/delivered?????  We have log files up to 26GB !!!!   with just these mesages in it !!!!

Question2:  Is it allowed to decrease the values of the config tags <page-size-bytes> and <max-size-bytes> while there are still messages in memory and in disk pages??????  (these settings were changed wile the server was stopped first)

We have experienced some strange behavior in the system like message counters that are not in sync with the actual nbr of messages (e.g. lost mesages) on the queue's (as far as we can see) . 

Queue's reporting messages count > 0 when ther are no messages to consume, when there should be messages to consume accoring to us!! (I started a discussion on that earlier)

To be honest I'm starting to have doubts in the integrity of the HornetQ implementation when it comes to handling paged messages. We see strange, unexplainable behaviour of HornetQ. That makes me very nervous!!!

Please help!

Regards,  Jan van de Klok