Try moving some of the mapped memory from the nodestore and relationshipstore to the propertystore, so they better match the size ratios of the files. Also increase the size of your Java heap. I think the program might be suffering from premature promotion. You can set cache_type=none if the GC pauses are still too high, but this will probably reduce the performance of the database, which may or may not be acceptable for you. You are configuring nodestore_mapped_memory_size twice, not that it matters much. I also recommend upgrading to 1.9.9 instead of 1.9.8, though it won’t make a difference for GC.
-- Chris Vest System Engineer, Neo Technology [ skype: mr.chrisvest, twitter: chvest ] > On 20 Mar 2015, at 08:56, S C Kannan <[email protected]> wrote: > > Hi All, > > We are Neo4j for last 2 years and currently we upgrade from 1.8.3 to > 1.9.8.Since we upgraded, we are seeing the follow info in messages.log file. > > 2015-03-20 04:37:54.242+0000 INFO [o.n.k.EmbeddedGraphDatabase]: GC Monitor: > Application threads blocked for an additional 400ms [total block time: 0.792s] > 2015-03-20 04:37:57.888+0000 INFO [o.n.k.EmbeddedGraphDatabase]: GC Monitor: > Application threads blocked for an additional 540ms [total block time: 1.332s] > 2015-03-20 04:38:07.293+0000 INFO [o.n.k.EmbeddedGraphDatabase]: GC Monitor: > Application threads blocked for an additional 475ms [total block time: 1.807s] > 2015-03-20 04:38:15.106+0000 INFO [o.n.k.EmbeddedGraphDatabase]: GC Monitor: > Application threads blocked for an additional 439ms [total block time: 2.246s] > > > We are using Embedded Neo4j DB. Our Neo4j Configuration file details are as > follows, > > node_keys_indexable=token,serialNumber,batchNo,promoId > node_auto_indexing=true > cache_type=gcr > nodestore_propertystore_mapped_memory_size=150M > nodestore_mapped_memory_size=100M > relationshipstore_mapped_memory_size=500M > strings_mapped_memory_size=150M > nodestore_mapped_memory_size=150M > relationship_auto_indexing=true > > Our Neo4j DB properties are as follows, > neostore.nodestore.db -- 18M > neostore.nodestore.db.id -- 1.2K > neostore.propertystore.db -- 316M > neostore.propertystore.db.arrays -- 128 > neostore.propertystore.db.arrays.id -- 9 > neostore.propertystore.db.id -- 57 > neostore.propertystore.db.index -- 4.2K > neostore.propertystore.db.index.id -- 9 > neostore.propertystore.db.index.keys -- 18K > neostore.propertystore.db.index.keys.id -- 9 > neostore.propertystore.db.strings -- 107M > neostore.propertystore.db.strings.id -- 17 > neostore.relationshipstore.db -- 252M > neostore.relationshipstore.db.id -- 321 > neostore.relationshiptypestore.db -- 795 > neostore.relationshiptypestore.db.id -- 9 > neostore.relationshiptypestore.db.names -- 6.0K > neostore.relationshiptypestore.db.names.id -- 9 > nioneo_logical.log.1 -- 1.3K > nioneo_logical.log.active -- 4 > > We have around 2M Nodes and 8M properties in our Database. We are strugged > for around 2 days and we have tried all options provided on forum. Kindly > help to sort it out. > > With Regards, > Kannan > > -- > You received this message because you are subscribed to the Google Groups > "Neo4j" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected] > <mailto:[email protected]>. > For more options, visit https://groups.google.com/d/optout > <https://groups.google.com/d/optout>. -- You received this message because you are subscribed to the Google Groups "Neo4j" 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.
