Neo4j权威指南第五章读书笔记

监控

截图

解释：

• Store Size: 存储容量， 调用操作系统获取Neo4j数据库存储文件的大小

  前面看过Neo4j数据的存储方式， 节点， 关系， 属性， 大段String， 索引都是分开存放的， 如图所示， 当前总容量是1.4GB

• IDAllocation: 分配的ID数

• PageCache 页面缓存 这个为什么木有数

• Transcations 事务

查询管理

开启超时查询保护：

dbms.transcation.timeout = 10s // 默认0表示不启用查询超时保护. 对java API访问时带的超时参数没影响

查询操作query(社区版没找到~~):

dbms.listQueries() 列出运行的语句

dbms.killQuery(queryId)

dbms.killQueries([queryId, queryId, ....])

数据收集器

dbms.udc.enabled=false 关闭数据收集(默认true)

安全管理

社区版的用户没有提供涉及角色、权限控制等安全管理功能。

只提供：

dbms.security.listUsers

dbms.security.showCurrentUsers

dbms.security.changePassword

dbms.security.createUser

dbms.security.deleteUser

运维优化

配置内存

./bin/neo4j-admin memrec 推荐配置

• 操作系统内存

  操作系统内存 = 可用内存 - (页面缓存 + 堆空间大小)

  index 和 schema 目录需要留出足够的内存用作操作系统的文件缓冲区， 否则索引文件不能完全装载在内存中，影响查询效率， 基本计算方式：

  系统内存 = 1GB + (graph.db/index) + graph.db / schema)

  graph.db 目前我们图数据库大概1.5G

• 页面缓存大小：

页面缓存用于缓存存储在磁盘上的Neo4j数据， 确保大部分数据缓存到内存中， 提高查询性能

dbms.memory.pagecache.size, 保证数据都能装进去， 再加20%预留

• 堆大小

堆内存大小足够大对维持并发操作是非常有用的， 一般建议 8~ 16GB之间

dbms.memory.heap.initial_size

dbms.memory.heap.max_size

设置为同一数值(如16000)， 以避免不必要的垃圾收集

• 堆的JVM调优：

  dbms.jvm.additional 附加JVM参数

        #**************************************************************
        # JVM Parameters
        #**********************************************************
        
        # G1GC generally strikes a good balance between throughput and tail
        # latency, without too much tuning.
        dbms.jvm.additional=-XX:+UseG1GC

        # Have common exceptions keep producing stack traces, so they can be
        # debugged regardless of how often logs are rotated.
        dbms.jvm.additional=-XX:-OmitStackTraceInFastThrow
        
        # Make sure that `initmemory` is not only allocated, but committed to
        # the process, before starting the database. This reduces memory
        # fragmentation, increasing the effectiveness of transparent huge
        # pages. It also reduces the possibility of seeing performance drop
        # due to heap-growing GC events, where a decrease in available page
        # cache leads to an increase in mean IO response time.
        # Try reducing the heap memory, if this flag degrades performance.
        dbms.jvm.additional=-XX:+AlwaysPreTouch
        
        # Trust that non-static final fields are really final.
        # This allows more optimizations and improves overall performance.
        # NOTE: Disable this if you use embedded mode, or have extensions or dependencies that may use reflection or
        # serialization to change the value of final fields!
        dbms.jvm.additional=-XX:+UnlockExperimentalVMOptions
        dbms.jvm.additional=-XX:+TrustFinalNonStaticFields
        
        # Disable explicit garbage collection, which is occasionally invoked by the JDK itself.
        dbms.jvm.additional=-XX:+DisableExplicitGC
        dbms.jvm.additional=-Xss100M
        # Disable explicit garbage collection, which is occasionally invoked by the JDK itself.
        dbms.jvm.additional=-XX:+DisableExplicitGC
        dbms.jvm.additional=-Xss100M
        
        # Remote JMX monitoring, uncomment and adjust the following lines as needed. Absolute paths to jmx.access and
        # jmx.password files are required.
        # Also make sure to update the jmx.access and jmx.password files with appropriate permission roles and passwords,
        # the shipped configuration contains only a read only role called 'monitor' with password 'Neo4j'.
        # For more details, see: http://download.oracle.com/javase/8/docs/technotes/guides/management/agent.html
        # On Unix based systems the jmx.password file needs to be owned by the user that will run the server,
        # and have permissions set to 0600.
        # For details on setting these file permissions on Windows see:
        #     http://docs.oracle.com/javase/8/docs/technotes/guides/management/security-windows.html
        #dbms.jvm.additional=-Dcom.sun.management.jmxremote.port=3637
        #dbms.jvm.additional=-Dcom.sun.management.jmxremote.authenticate=true
        #dbms.jvm.additional=-Dcom.sun.management.jmxremote.ssl=false
        #dbms.jvm.additional=-Dcom.sun.management.jmxremote.password.file=/absolute/path/to/conf/jmx.password
        #dbms.jvm.additional=-Dcom.sun.management.jmxremote.access.file=/absolute/path/to/conf/jmx.access
        
        # Some systems cannot discover host name automatically, and need this line configured:
        #dbms.jvm.additional=-Djava.rmi.server.hostname=$THE_NEO4J_SERVER_HOSTNAME
        
        # Expand Diffie Hellman (DH) key size from default 1024 to 2048 for DH-RSA cipher suites used in server TLS handshakes.
        # This is to protect the server from any potential passive eavesdropping.
        dbms.jvm.additional=-Djdk.tls.ephemeralDHKeySize=2048

        # This mitigates a DDoS vector.
        dbms.jvm.additional=-Djdk.tls.rejectClientInitiatedRenegotiation=true

(1) 分配新老两代大小比例：

   -XX:NewRation=N  # 年老代/年轻代 = N, N一般 2~ 8之间 大的年轻代适合更改排序大量数据， 运行并发线程也需要给年轻代更大空间，但年轻代过大容易引起频繁fullGC

(2) 堆的大小：

   最终目的是减少GC时间， 过大的堆会造成一般不GC， 一旦GC会耗费很长时间

(3) 并发垃圾回收：

   -XX:+UseG1GC

系统调优：

Neo4j读写许多读写操作， Linux默认CFQ公平排队调度IO. 期限调度器更适合数据库的特定IO工作负载情况， 优先读操作， 减少读的等待时间， 而增加写的等待时间。

修改驱动器sda：

echo 'deadline' > /sys/block/sda/queue/scheduler

cat  /sys/block/sda/queue/scheduler

磁盘内存：

快速固态硬盘 > 固态硬盘 > 机械硬盘

启动需要预热， 将索引加载到内存： CALL apoc.warmup.run(true);

dstat或vmstat等工具

备份恢复：

企业版提供了备份工具 neo4j-backup 备份到远程或本地, 支持全量， 增量

dbms.backup.enable=true # 默认值就是true
dbms.backup.address=<主机名/IP>:6362

社区版的只好自己写备份脚本了， 直接粘贴复制数据文件也可以， 但是官方建议更安全的数据导入导出工具(停服)

neo4j-admin dump --database=graph.db --to=${file}  # 备份压缩

neo4j-admin load --from=${file_catalog}/${file} --database=graph.db --force  # 恢复数据