Thanks MH, This is the basis for my dataset, which load it via neo4j-admin import, it take almost < 20 mins but take ~10 hours to using LOAD CSV
ID Allocation Node ID 143812726 Property ID 590958500 Relationship ID 244292244 Relationship Type ID 3 My LOAD CSV looks like that: 1. Create the CONSTRAINT and INDEX at the begin 2. Create the nodes USING PERIODIC COMMIT 50000 LOAD CSV FROM 'file:///XXXX.csv' AS row MERGE (c:Company { id: row[0] }) ON CREATE SET c.name = row[1], ...... ON MATCH SET c.name = row[1], ...... USING PERIODIC COMMIT 50000 LOAD CSV FROM 'file:///YYYY.csv' AS row MERGE (p:Employ { id: row[0] }) ON CREATE SET p.name = row[1], ...... ON MATCH SET p.name = row[1], ...... 3. Create the relations USING PERIODIC COMMIT 50000 LOAD CSV FROM 'file:///RRRR.csv' AS row MATCH (c:Company { id: row[0] }) MATCH (p:Employ { id: row[1] }) MERGE (c)-[r:EMPLOY]->(p) ON CREATE SET r.since=row[2], ...... ON MATCH SET r.since=row[2], ...... 在 2018年1月31日星期三 UTC+8下午8:08:36,Michael Hunger写道: > > It doesn't work like 1 or 2. > > You have to do (3). > > You have to completely clean out the other servers and have them copy from > master. > > Or you seed the whole cluster with the data from neo4j-admin import, i.e. > copy the database from master onto the other two. > Also make sure to use neo4j-shell on that data to create the necessary > indexes / cosntraints before using it to seed your cluster. > > How big is your data? And what does your LOAD CSV look like. > > Michael > > > Am 31.01.2018 um 05:01 schrieb Liping Huang <liping.h...@gmail.com > <javascript:>>: > > > > Hi there, > > > > Given I have a HA cluster with 3 servers: server1 - Master server2 - > Slave1 server3 - Slave2, as I want to keep all data up to date, and LOAD > CSV approach is too slow for my hurge data, > > > > So I want to load reload all data in offline with neo4j-admin import > functionality. > > > > Let's say I shutdown server1. ( so server2 will become the new Master > server ) > > > > then I load the new copy of data with neo4j-admin import functionality > in server1 > > > > then I start server1, as server2 already the Master in the HA cluster, > so server1 will join as the slaver > > > > so how about the sync behavior? > > 1. sync the new data from server1[Salver] to server2[Master] and > server3[Slaver] > > 2. sync the data from server2[Master] to server1[Slaver] > > 3. or I must delete the databases in server2, server3, and restart all > the servers, so server1 will become Master, server2 server3 will act as the > Slaver and sync all the data from Master > > > > please, thanks. > > > > > > -- > > 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 neo4j+un...@googlegroups.com <javascript:>. > > For more options, visit 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 neo4j+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.