Data loss with H2 HA / Clustered Mode

 

Hi,

We are trying to use the H2 DB after switching from Apache Derby.  This 
main reason for this was because of the clustering support in H2.  After a 
fair amount of testing we have now come across a problem with the clustered 
mode.

I have configured a system to use H2 (1.3.172) in clustered mode and access 
the DB using the supplied connection pool.  Our application has many 
threads accessing the DB concurrently (from Web Services and background 
tasks).  The main purpose of the DB is to hold a job queue where the job is 
added to the table, updated frequently while active and then set to 
complete.  I have found when the system is under load job records can 
“disappear” from the DB (many jobs being submitted and updates to existing 
records, ~40 out of 1000 jobs lost).  This doesn’t happen when the DB is 
run in standalone mode.

I don’t know how the clustering mode works internally but I was wondering 
if the transactions are not making it to one of the DB instances but this 
instance is then used for reading.  Would this be possible?  Could it be 
something I have set-up incorrectly?  Does clustering work OK with the 
connection pool class?  We had to switch to sequences for id generation, 
could I have done something wrong there?

Any help would be much appreciated.  I am happy to supply more details if 
required.

 

Regards,
Daniel Stone

-- 
You received this message because you are subscribed to the Google Groups "H2 
Database" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/h2-database.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to