I haven't used that yet but it'll be a great addition once the kinks are 
worked out.  Currently, I still use polls once a second, and then my server 
app handles all the other stuff.  Eventually, I will try switching to live 
query.  I have some classes where I poll by (datetime > lastprocessedTm). 
 I have others that are more demanding (@rid > lastRid).   This works great 
on a single node, and I'm happy to have confirmed that it can also run that 
way in distributed mode (increasing availability and persistence) as long 
as I use the default cluster.  

If I did need to add an additional counter, I was playing around with this, 
which while adding more complexity should have also worked (though I was 
getting an exception for some reason): 
https://github.com/orientechnologies/orientdb/issues/4316 - for a 
high-volume situation where maximizing sharding is helpful, that may still 
be a better approach (or this as well, however Yet Another Thing To 
Learn): http://doc.gridgain.org/latest/Distributed+AtomicSequence

But, as you can figure that synchronization would neutralize a good bit of 
the advantages of sharding in this usage case.

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to