[shared_buffers]的搜索结果

...ed-pipe shared-memory shared-memory-base-name=MYSQL The Pipe the MySQL Server will use socket=MYSQL The TCP/IP Port the MySQL Server will listen on port=3306 Path to installation directory. All paths are usually resolved relative to this. basedir="C:/Program Files/MySQL/MySQL Server 8.0/" Path to the database root datadir=C:/ProgramData/MySQL/MySQL Server 8.0/Data The default character set that will be used when a new schema or table is created and no character set is defined 创建新模式或表时使用的默认字符集，并且没有定义字符集 character-set-server= The default authentication plugin to be used when connecting to the server 连接到服务器时使用的默认身份验证插件 default_authentication_plugin=caching_sha2_password The default storage engine that will be used when create new tables when 当创建新表时将使用的默认存储引擎 default-storage-engine=INNODB Set the SQL mode to strict 将SQL模式设置为strict sql-mode="STRICT_TRANS_TABLES,NO_ENGINE_SUBSTITUTION" General and Slow logging. 一般和缓慢的日志。 log-output=NONE general-log=0 general_log_file="DESKTOP-NF9QETB.log" slow-query-log=0 slow_query_log_file="DESKTOP-NF9QETB-slow.log" long_query_time=10 Binary Logging. 二进制日志。 log-bin Error Logging. 错误日志记录。 log-error="DESKTOP-NF9QETB.err" Server Id. server-id=1 Indicates how table and database names are stored on disk and used in MySQL. 指示表名和数据库名如何存储在磁盘上并在MySQL中使用。 Value = 0: Table and database names are stored on disk using the lettercase specified in the CREATE TABLE or CREATE DATABASE statement. Name comparisons are case sensitive. You should not set this variable to 0 if you are running MySQL on a system that has case-insensitive file names (such as Windows or macOS). Value = 0:表名和数据库名使用CREATE Table或CREATE database语句中指定的lettercase存储在磁盘上。名称比较区分大小写。如果您在一个具有不区分大小写文件名(如Windows或macOS)的系统上运行MySQL，则不应将该变量设置为0。 Value = 1: Table names are stored in lowercase on disk and name comparisons are not case-sensitive. MySQL converts all table names to lowercase on storage and lookup. This behavior also applies to database names and table aliases. 表名以小写存储在磁盘上，并且名称比较不区分大小写。MySQL在存储和查找时将所有表名转换为小写。此行为也适用于数据库名称和表别名。 Value = 3, Table and database names are stored on disk using the lettercase specified in the CREATE TABLE or CREATE DATABASE statement, but MySQL converts them to lowercase on lookup. Name comparisons are not case sensitive. This works only on file systems that are not case-sensitive! InnoDB table names and view names are stored in lowercase, as for Value = 1.表名和数据库名使用CREATE Table或CREATE database语句中指定的lettercase存储在磁盘上，但是MySQL在查找时将它们转换为小写。名称比较不区分大小写。这只适用于不区分大小写的文件系统!InnoDB表名和视图名以小写存储，Value = 1。 NOTE: lower_case_table_names can only be configured when initializing the server. Changing the lower_case_table_names setting after the server is initialized is prohibited. lower_case_table_names=1 Secure File Priv. 权限安全文件 secure-file-priv="C:/ProgramData/MySQL/MySQL Server 8.0/Uploads" The maximum amount of concurrent sessions the MySQL server will allow. One of these connections will be reserved for a user with SUPER privileges to allow the administrator to login even if the connection limit has been reached. MySQL服务器允许的最大并发会话量。这些连接中的一个将保留给具有超级特权的用户，以便允许管理员登录，即使已经达到连接限制。 max_connections=151 The number of open tables for all threads. Increasing this value increases the number of file descriptors that mysqld requires. Therefore you have to make sure to set the amount of open files allowed to at least 4096 in the variable "open-files-limit" in 为所有线程打开的表的数量。增加这个值会增加mysqld需要的文件描述符的数量。因此，您必须确保在[mysqld_safe]节中的变量“open-files-limit”中将允许打开的文件数量至少设置为4096 section [mysqld_safe] table_open_cache=2000 Maximum size for internal (in-memory) temporary tables. If a table grows larger than this value, it is automatically converted to disk based table This limitation is for a single table. There can be many of them. 内部(内存)临时表的最大大小。如果一个表比这个值大，那么它将自动转换为基于磁盘的表。可以有很多。 tmp_table_size=94M How many threads we should keep in a cache for reuse. When a client disconnects, the client's threads are put in the cache if there aren't more than thread_cache_size threads from before. This greatly reduces the amount of thread creations needed if you have a lot of new connections. (Normally this doesn't give a notable performance improvement if you have a good thread implementation.) 我们应该在缓存中保留多少线程以供重用。当客户机断开连接时，如果之前的线程数不超过thread_cache_size，则将客户机的线程放入缓存。如果您有很多新连接，这将大大减少所需的线程创建量(通常，如果您有一个良好的线程实现，这不会带来显著的性能改进)。 thread_cache_size=10 MyISAM Specific options The maximum size of the temporary file MySQL is allowed to use while recreating the index (during REPAIR, ALTER TABLE or LOAD DATA INFILE. If the file-size would be bigger than this, the index will be created through the key cache (which is slower). MySQL允许在重新创建索引时(在修复、修改表或加载数据时)使用临时文件的最大大小。如果文件大小大于这个值，那么索引将通过键缓存创建(这比较慢)。 myisam_max_sort_file_size=100G If the temporary file used for fast index creation would be bigger than using the key cache by the amount specified here, then prefer the key cache method. This is mainly used to force long character keys in large tables to use the slower key cache method to create the index. myisam_sort_buffer_size=179M Size of the Key Buffer, used to cache index blocks for MyISAM tables. Do not set it larger than 30% of your available memory, as some memory is also required by the OS to cache rows. Even if you're not using MyISAM tables, you should still set it to 8-64M as it will also be used for internal temporary disk tables. 如果用于快速创建索引的临时文件比这里指定的使用键缓存的文件大，则首选键缓存方法。这主要用于强制大型表中的长字符键使用较慢的键缓存方法来创建索引。 key_buffer_size=8M Size of the buffer used for doing full table scans of MyISAM tables. Allocated per thread, if a full scan is needed. 用于对MyISAM表执行全表扫描的缓冲区的大小。如果需要完整的扫描，则为每个线程分配。 read_buffer_size=256K read_rnd_buffer_size=512K INNODB Specific options INNODB特定选项 innodb_data_home_dir= Use this option if you have a MySQL server with InnoDB support enabled but you do not plan to use it. This will save memory and disk space and speed up some things. 如果您启用了一个支持InnoDB的MySQL服务器，但是您不打算使用它，那么可以使用这个选项。这将节省内存和磁盘空间，并加快一些事情。skip-innodb skip-innodb If set to 1, InnoDB will flush (fsync) the transaction logs to the disk at each commit, which offers full ACID behavior. If you are willing to compromise this safety, and you are running small transactions, you may set this to 0 or 2 to reduce disk I/O to the logs. Value 0 means that the log is only written to the log file and the log file flushed to disk approximately once per second. Value 2 means the log is written to the log file at each commit, but the log file is only flushed to disk approximately once per second. 如果设置为1,InnoDB将在每次提交时将事务日志刷新(fsync)到磁盘，这将提供完整的ACID行为。如果您愿意牺牲这种安全性，并且正在运行小型事务，您可以将其设置为0或2，以将磁盘I/O减少到日志。值0表示日志仅写入日志文件，日志文件大约每秒刷新一次磁盘。值2表示日志在每次提交时写入日志文件，但是日志文件大约每秒只刷新一次磁盘。 innodb_flush_log_at_trx_commit=1 The size of the buffer InnoDB uses for buffering log data. As soon as it is full, InnoDB will have to flush it to disk. As it is flushed once per second anyway, it does not make sense to have it very large (even with long transactions).InnoDB用于缓冲日志数据的缓冲区大小。一旦它满了，InnoDB就必须将它刷新到磁盘。由于它无论如何每秒刷新一次，所以将它设置为非常大的值是没有意义的(即使是长事务)。 innodb_log_buffer_size=5M InnoDB, unlike MyISAM, uses a buffer pool to cache both indexes and row data. The bigger you set this the less disk I/O is needed to access data in tables. On a dedicated database server you may set this parameter up to 80% of the machine physical memory size. Do not set it too large, though, because competition of the physical memory may cause paging in the operating system. Note that on 32bit systems you might be limited to 2-3.5G of user level memory per process, so do not set it too high. 与MyISAM不同，InnoDB使用缓冲池来缓存索引和行数据。设置的值越大，访问表中的数据所需的磁盘I/O就越少。在专用数据库服务器上，可以将该参数设置为机器物理内存大小的80%。但是，不要将它设置得太大，因为物理内存的竞争可能会导致操作系统中的分页。注意，在32位系统上，每个进程的用户级内存可能被限制在2-3.5G，所以不要设置得太高。 innodb_buffer_pool_size=20M Size of each log file in a log group. You should set the combined size of log files to about 25%-100% of your buffer pool size to avoid unneeded buffer pool flush activity on log file overwrite. However, note that a larger logfile size will increase the time needed for the recovery process. 日志组中每个日志文件的大小。您应该将日志文件的合并大小设置为缓冲池大小的25%-100%，以避免在覆盖日志文件时出现不必要的缓冲池刷新活动。但是，请注意，较大的日志文件大小将增加恢复过程所需的时间。 innodb_log_file_size=48M Number of threads allowed inside the InnoDB kernel. The optimal value depends highly on the application, hardware as well as the OS scheduler properties. A too high value may lead to thread thrashing. InnoDB内核中允许的线程数。最优值在很大程度上取决于应用程序、硬件以及OS调度程序属性。过高的值可能导致线程抖动。 innodb_thread_concurrency=9 The increment size (in MB) for extending the size of an auto-extend InnoDB system tablespace file when it becomes full. 增量大小(以MB为单位)，用于在表空间满时扩展自动扩展的InnoDB系统表空间文件的大小。 innodb_autoextend_increment=128 The number of regions that the InnoDB buffer pool is divided into. For systems with buffer pools in the multi-gigabyte range, dividing the buffer pool into separate instances can improve concurrency, by reducing contention as different threads read and write to cached pages. InnoDB缓冲池划分的区域数。对于具有多gb缓冲池的系统，将缓冲池划分为单独的实例可以提高并发性，因为不同的线程对缓存页面的读写会减少争用。 innodb_buffer_pool_instances=8 Determines the number of threads that can enter InnoDB concurrently. 确定可以同时进入InnoDB的线程数 innodb_concurrency_tickets=5000 Specifies how long in milliseconds (ms) a block inserted into the old sublist must stay there after its first access before it can be moved to the new sublist. 指定插入到旧子列表中的块必须在第一次访问之后停留多长时间(毫秒)，然后才能移动到新子列表。 innodb_old_blocks_time=1000 It specifies the maximum number of .ibd files that MySQL can keep open at one time. The minimum value is 10. 它指定MySQL一次可以打开的.ibd文件的最大数量。最小值是10。 innodb_open_files=300 When this variable is enabled, InnoDB updates statistics during metadata statements. 当启用此变量时，InnoDB会在元数据语句期间更新统计信息。 innodb_stats_on_metadata=0 When innodb_file_per_table is enabled (the default in 5.6.6 and higher), InnoDB stores the data and indexes for each newly created table in a separate .ibd file, rather than in the system tablespace. 当启用innodb_file_per_table(5.6.6或更高版本的默认值)时，InnoDB将每个新创建的表的数据和索引存储在单独的.ibd文件中，而不是系统表空间中。 innodb_file_per_table=1 Use the following list of values: 0 for crc32, 1 for strict_crc32, 2 for innodb, 3 for strict_innodb, 4 for none, 5 for strict_none. 使用以下值列表:0表示crc32, 1表示strict_crc32, 2表示innodb, 3表示strict_innodb, 4表示none, 5表示strict_none。 innodb_checksum_algorithm=0 The number of outstanding connection requests MySQL can have. This option is useful when the main MySQL thread gets many connection requests in a very short time. It then takes some time (although very little) for the main thread to check the connection and start a new thread. The back_log value indicates how many requests can be stacked during this short time before MySQL momentarily stops answering new requests. You need to increase this only if you expect a large number of connections in a short period of time. MySQL可以有多少未完成连接请求。当MySQL主线程在很短的时间内收到许多连接请求时，这个选项非常有用。然后，主线程需要一些时间(尽管很少)来检查连接并启动一个新线程。back_log值表示在MySQL暂时停止响应新请求之前的短时间内可以堆多少个请求。只有当您预期在短时间内会有大量连接时，才需要增加这个值。 back_log=80 If this is set to a nonzero value, all tables are closed every flush_time seconds to free up resources and synchronize unflushed data to disk. This option is best used only on systems with minimal resources. 如果将该值设置为非零值，则每隔flush_time秒关闭所有表，以释放资源并将未刷新的数据同步到磁盘。这个选项最好只在资源最少的系统上使用。 flush_time=0 The minimum size of the buffer that is used for plain index scans, range index scans, and joins that do not use 用于普通索引扫描、范围索引扫描和不使用索引执行全表扫描的连接的缓冲区的最小大小。 indexes and thus perform full table scans. join_buffer_size=200M The maximum size of one packet or any generated or intermediate string, or any parameter sent by the mysql_stmt_send_long_data() C API function. 由mysql_stmt_send_long_data() C API函数发送的一个包或任何生成的或中间字符串或任何参数的最大大小 max_allowed_packet=500M If more than this many successive connection requests from a host are interrupted without a successful connection, the server blocks that host from performing further connections. 如果在没有成功连接的情况下中断了来自主机的多个连续连接请求，则服务器将阻止主机执行进一步的连接。 max_connect_errors=100 Changes the number of file descriptors available to mysqld. You should try increasing the value of this option if mysqld gives you the error "Too many open files". 更改mysqld可用的文件描述符的数量。如果mysqld给您的错误是“打开的文件太多”，您应该尝试增加这个选项的值。 open_files_limit=4161 If you see many sort_merge_passes per second in SHOW GLOBAL STATUS output, you can consider increasing the sort_buffer_size value to speed up ORDER BY or GROUP BY operations that cannot be improved with query optimization or improved indexing. 如果在SHOW GLOBAL STATUS输出中每秒看到许多sort_merge_passes，可以考虑增加sort_buffer_size值，以加快ORDER BY或GROUP BY操作的速度，这些操作无法通过查询优化或改进索引来改进。 sort_buffer_size=1M The number of table definitions (from .frm files) that can be stored in the definition cache. If you use a large number of tables, you can create a large table definition cache to speed up opening of tables. The table definition cache takes less space and does not use file descriptors, unlike the normal table cache. The minimum and default values are both 400. 可以存储在定义缓存中的表定义的数量(来自.frm文件)。如果使用大量表，可以创建一个大型表定义缓存来加速表的打开。与普通的表缓存不同，表定义缓存占用更少的空间，并且不使用文件描述符。最小值和默认值都是400。 table_definition_cache=1400 Specify the maximum size of a row-based binary log event, in bytes. Rows are grouped into events smaller than this size if possible. The value should be a multiple of 256. 指定基于行的二进制日志事件的最大大小，单位为字节。如果可能，将行分组为小于此大小的事件。这个值应该是256的倍数。 binlog_row_event_max_size=8K If the value of this variable is greater than 0, a replication slave synchronizes its master.info file to disk. (using fdatasync()) after every sync_master_info events. 如果该变量的值大于0，则复制奴隶将其主.info文件同步到磁盘。(在每个sync_master_info事件之后使用fdatasync())。 sync_master_info=10000 If the value of this variable is greater than 0, the MySQL server synchronizes its relay log to disk. (using fdatasync()) after every sync_relay_log writes to the relay log. 如果这个变量的值大于0,MySQL服务器将其中继日志同步到磁盘。(在每个sync_relay_log写入到中继日志之后使用fdatasync())。 sync_relay_log=10000 If the value of this variable is greater than 0, a replication slave synchronizes its relay-log.info file to disk. (using fdatasync()) after every sync_relay_log_info transactions. 如果该变量的值大于0，则复制奴隶将其中继日志.info文件同步到磁盘。(在每个sync_relay_log_info事务之后使用fdatasync())。 sync_relay_log_info=10000 Load mysql plugins at start."plugin_x ; plugin_y". 开始时加载mysql插件。“plugin_x;plugin_y” plugin_load The TCP/IP Port the MySQL Server X Protocol will listen on. MySQL服务器X协议将监听TCP/IP端口。 loose_mysqlx_port=33060 本篇文章为转载内容。原文链接：https://blog.csdn.net/mywpython/article/details/89499852。 该文由互联网用户投稿提供，文中观点代表作者本人意见，并不代表本站的立场。 作为信息平台，本站仅提供文章转载服务，并不拥有其所有权，也不对文章内容的真实性、准确性和合法性承担责任。 如发现本文存在侵权、违法、违规或事实不符的情况，请及时联系我们，我们将第一时间进行核实并删除相应内容。

...响 2.1 shared_buffers设置不合理 shared_buffers是PostgreSQL用于缓存数据的重要参数，其大小直接影响到数据库的查询性能。要是你把这数值设得过小，就等于是在让磁盘I/O忙个不停，频繁操作起来，就像个永不停歇的陀螺，会拖累整体性能，让系统跑得像只乌龟。反过来，如果你一不留神把数值调得过大，那就像是在内存里开辟了一大片空地却闲置不用，这就白白浪费了宝贵的内存资源，还会把其他系统进程挤得没地方住，人家也会闹情绪的。 postgresql -- 在postgresql.conf中调整shared_buffers值 shared_buffers = 4GB -- 假设服务器有足够内存支持此设置 2.2 work_mem不足 work_mem定义了每个SQL查询可以使用的内存量，对于复杂的排序、哈希操作等至关重要。过低的work_mem设定可能导致大量临时文件生成，进一步降低性能。 postgresql -- 调整work_mem大小 work_mem = 64MB -- 根据实际业务负载进行合理调整 3. 配置失误导致的故障案例 3.1 max_connections设置过高 max_connections参数限制了PostgreSQL同时接受的最大连接数。如果设置得过高，却没考虑服务器的实际承受能力，就像让一个普通人硬扛大铁锤，早晚得累垮。这样一来，系统资源就会被消耗殆尽，好比车票都被抢光了，新的连接请求就无法挤上这趟“网络列车”。最终，整个系统可能就要“罢工”瘫痪啦。 postgresql -- 不合理的高连接数设置示例 max_connections = 500 -- 若服务器硬件条件不足以支撑如此多的并发连接，则可能引发故障 3.2 日志设置不当造成磁盘空间耗尽 log_line_prefix、log_directory等日志相关参数设置不当，可能导致日志文件迅速增长，占用过多磁盘空间，进而引发数据库服务停止。 postgresql -- 错误的日志设置示例 log_line_prefix = '%t [%p]: ' -- 时间戳和进程ID前缀可能会使日志行变得冗长 log_directory = '/var/log/postgresql' -- 如果不加以定期清理，日志文件可能会撑满整个分区 4. 探讨与建议 面对PostgreSQL的系统配置问题，我们需要深入了解每个参数的含义以及它们在不同场景下的最佳实践。优化配置是一个持续的过程，需要结合业务特性和硬件资源来进行细致调优。 - 理解需求：首先，应了解业务特点，包括数据量大小、查询复杂度、并发访问量等因素。 - 监控分析：借助pg_stat_activity、pg_stat_bgwriter等视图监控数据库运行状态，结合如pgBadger、pg_top等工具分析性能瓶颈。 - 逐步调整：每次只更改一个参数，观察并评估效果，切忌盲目跟从网络上的推荐配置。 总结来说，PostgreSQL的强大性能背后，合理的配置是关键。要让咱们的数据库系统跑得溜又稳，像老黄牛一样可靠，给业务发展扎扎实实当好坚强后盾，那就必须把这些参数整得门儿清，调校得恰到好处才行。

...g default shared_buffers ... 128MBselecting dynamic shared memory implementation ... posixcreating configuration files ... okrunning bootstrap script ... okperforming post-bootstrap initialization ... creating cluster information ... oksyncing data to disk ... okfreezing database template0 ... okfreezing database template1 ... okfreezing database postgres ... okWARNING: enabling "trust" authentication for local connectionsYou can change this by editing pg_hba.conf or using the option -A, or--auth-local and --auth-host, the next time you run initdb.Success.Done.Starting coordinator master.Starting coordinator master coord1ERROR: target coordinator master coord1 is already running now. Skip initialization.Starting coordinator master coord22019-05-30 21:09:25.562 EDT [2148] LOG: listening on IPv4 address "0.0.0.0", port 54322019-05-30 21:09:25.562 EDT [2148] LOG: listening on IPv6 address "::", port 54322019-05-30 21:09:25.563 EDT [2148] LOG: listening on Unix socket "/tmp/.s.PGSQL.5432"2019-05-30 21:09:25.601 EDT [2149] LOG: database system was shut down at 2019-05-30 21:09:22 EDT2019-05-30 21:09:25.605 EDT [2148] LOG: database system is ready to accept connections2019-05-30 21:09:25.612 EDT [2156] LOG: cluster monitor startedDone.Initialize all the datanode masters.Initialize the datanode master datanode1.Initialize the datanode master datanode2.The files belonging to this database system will be owned by user "postgres".This user must also own the server process.The database cluster will be initialized with locale "en_US.UTF-8".The default database encoding has accordingly been set to "UTF8".The default text search configuration will be set to "english".Data page checksums are disabled.fixing permissions on existing directory /home/postgres/pgxc/nodes/datanode/datanode1 ... okcreating subdirectories ... okselecting default max_connections ... 100selecting default shared_buffers ... 128MBselecting dynamic shared memory implementation ... posixcreating configuration files ... okrunning bootstrap script ... okperforming post-bootstrap initialization ... creating cluster information ... oksyncing data to disk ... okfreezing database template0 ... okfreezing database template1 ... okfreezing database postgres ... okWARNING: enabling "trust" authentication for local connectionsYou can change this by editing pg_hba.conf or using the option -A, or--auth-local and --auth-host, the next time you run initdb.Success.The files belonging to this database system will be owned by user "postgres".This user must also own the server process.The database cluster will be initialized with locale "en_US.UTF-8".The default database encoding has accordingly been set to "UTF8".The default text search configuration will be set to "english".Data page checksums are disabled.fixing permissions on existing directory /home/postgres/pgxc/nodes/datanode/datanode2 ... okcreating subdirectories ... okselecting default max_connections ... 100selecting default shared_buffers ... 128MBselecting dynamic shared memory implementation ... posixcreating configuration files ... okrunning bootstrap script ... okperforming post-bootstrap initialization ... creating cluster information ... oksyncing data to disk ... okfreezing database template0 ... okfreezing database template1 ... okfreezing database postgres ... okWARNING: enabling "trust" authentication for local connectionsYou can change this by editing pg_hba.conf or using the option -A, or--auth-local and --auth-host, the next time you run initdb.Success.Done.Starting all the datanode masters.Starting datanode master datanode1.WARNING: datanode master datanode1 is running now. Skipping.Starting datanode master datanode2.2019-05-30 21:09:33.352 EDT [2404] LOG: listening on IPv4 address "0.0.0.0", port 154322019-05-30 21:09:33.352 EDT [2404] LOG: listening on IPv6 address "::", port 154322019-05-30 21:09:33.355 EDT [2404] LOG: listening on Unix socket "/tmp/.s.PGSQL.15432"2019-05-30 21:09:33.392 EDT [2404] LOG: redirecting log output to logging collector process2019-05-30 21:09:33.392 EDT [2404] HINT: Future log output will appear in directory "pg_log".Done.psql: FATAL: no pg_hba.conf entry for host "192.168.20.132", user "postgres", database "postgres"psql: FATAL: no pg_hba.conf entry for host "192.168.20.132", user "postgres", database "postgres"Done.psql: FATAL: no pg_hba.conf entry for host "192.168.20.132", user "postgres", database "postgres"psql: FATAL: no pg_hba.conf entry for host "192.168.20.132", user "postgres", database "postgres"Done.[postgres@gtm ~]$ pgxc_ctl -c /home/postgres/pgxc_ctl/pgxc_ctl.conf stop all/bin/bashInstalling pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Installing pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Reading configuration using /home/postgres/pgxc_ctl/pgxc_ctl_bash --home /home/postgres/pgxc_ctl --configuration /home/postgres/pgxc_ctl/pgxc_ctl.conf/home/postgres/pgxc_ctl/pgxc_ctl.conf: line 189: $coordExtraConfig: ambiguous redirectFinished reading configuration. PGXC_CTL START Current directory: /home/postgres/pgxc_ctlStopping all the coordinator masters.Stopping coordinator master coord1.Stopping coordinator master coord2.pg_ctl: directory "/home/postgres/pgxc/nodes/coord/coord1" does not existDone.Stopping all the datanode masters.Stopping datanode master datanode1.Stopping datanode master datanode2.pg_ctl: PID file "/home/postgres/pgxc/nodes/datanode/datanode1/postmaster.pid" does not existIs server running?Done.Stop GTM masterwaiting for server to shut down.... doneserver stopped[postgres@gtm ~]$ pgxc_ctl/bin/bashInstalling pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Installing pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Reading configuration using /home/postgres/pgxc_ctl/pgxc_ctl_bash --home /home/postgres/pgxc_ctl --configuration /home/postgres/pgxc_ctl/pgxc_ctl.conf/home/postgres/pgxc_ctl/pgxc_ctl.conf: line 189: $coordExtraConfig: ambiguous redirectFinished reading configuration. PGXC_CTL START Current directory: /home/postgres/pgxc_ctlPGXC monitor allNot running: gtm masterRunning: coordinator master coord1Not running: coordinator master coord2Running: datanode master datanode1Not running: datanode master datanode2PGXC stop coordinator master coord1Stopping coordinator master coord1.pg_ctl: directory "/home/postgres/pgxc/nodes/coord/coord1" does not existDone.PGXC stop datanode master datanode1Stopping datanode master datanode1.pg_ctl: PID file "/home/postgres/pgxc/nodes/datanode/datanode1/postmaster.pid" does not existIs server running?Done.PGXC monitor allNot running: gtm masterRunning: coordinator master coord1Not running: coordinator master coord2Running: datanode master datanode1Not running: datanode master datanode2PGXC monitor allNot running: gtm masterNot running: coordinator master coord1Not running: coordinator master coord2Not running: datanode master datanode1Not running: datanode master datanode2PGXC exit[postgres@gtm ~]$ pgxc_ctl -c /home/postgres/pgxc_ctl/pgxc_ctl.conf init all/bin/bashInstalling pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Installing pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Reading configuration using /home/postgres/pgxc_ctl/pgxc_ctl_bash --home /home/postgres/pgxc_ctl --configuration /home/postgres/pgxc_ctl/pgxc_ctl.conf/home/postgres/pgxc_ctl/pgxc_ctl.conf: line 189: $coordExtraConfig: ambiguous redirectFinished reading configuration. PGXC_CTL START Current directory: /home/postgres/pgxc_ctlInitialize GTM masterERROR: target directory (/home/postgres/pgxc/nodes/gtm) exists and not empty. Skip GTM initilializationDone.Start GTM masterserver startingInitialize all the coordinator masters.Initialize coordinator master coord1.Initialize coordinator master coord2.The files belonging to this database system will be owned by user "postgres".This user must also own the server process.The database cluster will be initialized with locale "en_US.UTF-8".The default database encoding has accordingly been set to "UTF8".The default text search configuration will be set to "english".Data page checksums are disabled.fixing permissions on existing directory /home/postgres/pgxc/nodes/coord/coord1 ... okcreating subdirectories ... okselecting default max_connections ... 100selecting default shared_buffers ... 128MBselecting dynamic shared memory implementation ... posixcreating configuration files ... okrunning bootstrap script ... okperforming post-bootstrap initialization ... creating cluster information ... oksyncing data to disk ... okfreezing database template0 ... okfreezing database template1 ... okfreezing database postgres ... okWARNING: enabling "trust" authentication for local connectionsYou can change this by editing pg_hba.conf or using the option -A, or--auth-local and --auth-host, the next time you run initdb.Success.The files belonging to this database system will be owned by user "postgres".This user must also own the server process.The database cluster will be initialized with locale "en_US.UTF-8".The default database encoding has accordingly been set to "UTF8".The default text search configuration will be set to "english".Data page checksums are disabled.fixing permissions on existing directory /home/postgres/pgxc/nodes/coord/coord2 ... okcreating subdirectories ... okselecting default max_connections ... 100selecting default shared_buffers ... 128MBselecting dynamic shared memory implementation ... posixcreating configuration files ... okrunning bootstrap script ... okperforming post-bootstrap initialization ... creating cluster information ... oksyncing data to disk ... okfreezing database template0 ... okfreezing database template1 ... okfreezing database postgres ... okWARNING: enabling "trust" authentication for local connectionsYou can change this by editing pg_hba.conf or using the option -A, or--auth-local and --auth-host, the next time you run initdb.Success.Done.Starting coordinator master.Starting coordinator master coord1Starting coordinator master coord22019-05-30 21:13:03.998 EDT [25137] LOG: listening on IPv4 address "0.0.0.0", port 54322019-05-30 21:13:03.998 EDT [25137] LOG: listening on IPv6 address "::", port 54322019-05-30 21:13:04.000 EDT [25137] LOG: listening on Unix socket "/tmp/.s.PGSQL.5432"2019-05-30 21:13:04.038 EDT [25138] LOG: database system was shut down at 2019-05-30 21:13:00 EDT2019-05-30 21:13:04.042 EDT [25137] LOG: database system is ready to accept connections2019-05-30 21:13:04.049 EDT [25145] LOG: cluster monitor started2019-05-30 21:13:04.020 EDT [2730] LOG: listening on IPv4 address "0.0.0.0", port 54322019-05-30 21:13:04.020 EDT [2730] LOG: listening on IPv6 address "::", port 54322019-05-30 21:13:04.021 EDT [2730] LOG: listening on Unix socket "/tmp/.s.PGSQL.5432"2019-05-30 21:13:04.057 EDT [2731] LOG: database system was shut down at 2019-05-30 21:13:00 EDT2019-05-30 21:13:04.061 EDT [2730] LOG: database system is ready to accept connections2019-05-30 21:13:04.062 EDT [2738] LOG: cluster monitor startedDone.Initialize all the datanode masters.Initialize the datanode master datanode1.Initialize the datanode master datanode2.The files belonging to this database system will be owned by user "postgres".This user must also own the server process.The database cluster will be initialized with locale "en_US.UTF-8".The default database encoding has accordingly been set to "UTF8".The default text search configuration will be set to "english".Data page checksums are disabled.fixing permissions on existing directory /home/postgres/pgxc/nodes/datanode/datanode1 ... okcreating subdirectories ... okselecting default max_connections ... 100selecting default shared_buffers ... 128MBselecting dynamic shared memory implementation ... posixcreating configuration files ... okrunning bootstrap script ... okperforming post-bootstrap initialization ... creating cluster information ... oksyncing data to disk ... okfreezing database template0 ... okfreezing database template1 ... okfreezing database postgres ... okWARNING: enabling "trust" authentication for local connectionsYou can change this by editing pg_hba.conf or using the option -A, or--auth-local and --auth-host, the next time you run initdb.Success.The files belonging to this database system will be owned by user "postgres".This user must also own the server process.The database cluster will be initialized with locale "en_US.UTF-8".The default database encoding has accordingly been set to "UTF8".The default text search configuration will be set to "english".Data page checksums are disabled.fixing permissions on existing directory /home/postgres/pgxc/nodes/datanode/datanode2 ... okcreating subdirectories ... okselecting default max_connections ... 100selecting default shared_buffers ... 128MBselecting dynamic shared memory implementation ... posixcreating configuration files ... okrunning bootstrap script ... okperforming post-bootstrap initialization ... creating cluster information ... oksyncing data to disk ... okfreezing database template0 ... okfreezing database template1 ... okfreezing database postgres ... okWARNING: enabling "trust" authentication for local connectionsYou can change this by editing pg_hba.conf or using the option -A, or--auth-local and --auth-host, the next time you run initdb.Success.Done.Starting all the datanode masters.Starting datanode master datanode1.Starting datanode master datanode2.2019-05-30 21:13:12.077 EDT [25392] LOG: listening on IPv4 address "0.0.0.0", port 154322019-05-30 21:13:12.077 EDT [25392] LOG: listening on IPv6 address "::", port 154322019-05-30 21:13:12.079 EDT [25392] LOG: listening on Unix socket "/tmp/.s.PGSQL.15432"2019-05-30 21:13:12.114 EDT [25392] LOG: redirecting log output to logging collector process2019-05-30 21:13:12.114 EDT [25392] HINT: Future log output will appear in directory "pg_log".2019-05-30 21:13:12.079 EDT [2985] LOG: listening on IPv4 address "0.0.0.0", port 154322019-05-30 21:13:12.079 EDT [2985] LOG: listening on IPv6 address "::", port 154322019-05-30 21:13:12.081 EDT [2985] LOG: listening on Unix socket "/tmp/.s.PGSQL.15432"2019-05-30 21:13:12.117 EDT [2985] LOG: redirecting log output to logging collector process2019-05-30 21:13:12.117 EDT [2985] HINT: Future log output will appear in directory "pg_log".Done.psql: FATAL: no pg_hba.conf entry for host "192.168.20.132", user "postgres", database "postgres"psql: FATAL: no pg_hba.conf entry for host "192.168.20.132", user "postgres", database "postgres"Done.psql: FATAL: no pg_hba.conf entry for host "192.168.20.132", user "postgres", database "postgres"psql: FATAL: no pg_hba.conf entry for host "192.168.20.132", user "postgres", database "postgres"Done. 启动 pgxc_ctl -c /home/postgres/pgxc_ctl/pgxc_ctl.conf start all 关闭 pgxc_ctl -c /home/postgres/pgxc_ctl/pgxc_ctl.conf stop all 查看集群状态 [postgres@gtm ~]$ pgxc_ctl monitor all/bin/bashInstalling pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Installing pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Reading configuration using /home/postgres/pgxc_ctl/pgxc_ctl_bash --home /home/postgres/pgxc_ctl --configuration /home/postgres/pgxc_ctl/pgxc_ctl.conf/home/postgres/pgxc_ctl/pgxc_ctl.conf: line 189: $coordExtraConfig: ambiguous redirectFinished reading configuration. PGXC_CTL START Current directory: /home/postgres/pgxc_ctlRunning: gtm masterRunning: coordinator master coord1Running: coordinator master coord2Running: datanode master datanode1Running: datanode master datanode2 配置集群信息 分别在数据节点、协调器节点上分别执行以下命令： 注：本节点只执行修改操作即可(alert node)，其他节点执行创建命令(create node)。因为本节点已经包含本节点的信息。 create node coord1 with (type=coordinator,host=xl1, port=5432);create node coord2 with (type=coordinator,host=xl2, port=5432);alter node coord1 with (type=coordinator,host=xl1, port=5432);alter node coord2 with (type=coordinator,host=xl2, port=5432);create node datanode1 with (type=datanode, host=xl1,port=15432,primary=true,PREFERRED);create node datanode2 with (type=datanode, host=xl2,port=15432);alter node datanode1 with (type=datanode, host=xl1,port=15432,primary=true,PREFERRED);alter node datanode2 with (type=datanode, host=xl2,port=15432);select pgxc_pool_reload(); 分别登陆数据节点、协调器节点验证 postgres= select from pgxc_node;node_name | node_type | node_port | node_host | nodeis_primary | nodeis_preferred | node_id-----------+-----------+-----------+-----------+----------------+------------------+-------------coord1 | C | 5432 | xl1 | f | f | 1885696643coord2 | C | 5432 | xl2 | f | f | -1197102633datanode2 | D | 15432 | xl2 | f | f | -905831925datanode1 | D | 15432 | xl1 | t | f | 888802358(4 rows) 测试 插入数据 在数据节点1，执行相关操作。 通过协调器端口登录PG [postgres@xl1 ~]$ psql -p 5432psql (PGXL 10r1.1, based on PG 10.6 (Postgres-XL 10r1.1))Type "help" for help.postgres= create database lei;CREATE DATABASEpostgres= \c lei;You are now connected to database "lei" as user "postgres".lei= create table test1(id int,name text);CREATE TABLElei= insert into test1(id,name) select generate_series(1,8),'测试';INSERT 0 8lei= select from test1;id | name----+------1 | 测试2 | 测试5 | 测试6 | 测试8 | 测试3 | 测试4 | 测试7 | 测试(8 rows) 注：默认创建的表为分布式表，也就是每个数据节点值存储表的部分数据。关于表类型具体说明，下面有说明。 通过15432端口登录数据节点，查看数据 有5条数据 [postgres@xl1 ~]$ psql -p 15432psql (PGXL 10r1.1, based on PG 10.6 (Postgres-XL 10r1.1))Type "help" for help.postgres= \c lei;You are now connected to database "lei" as user "postgres".lei= select from test1;id | name----+------1 | 测试2 | 测试5 | 测试6 | 测试8 | 测试(5 rows) 登录到节点2，查看数据 有3条数据 [postgres@xl2 ~]$ psql -p15432psql (PGXL 10r1.1, based on PG 10.6 (Postgres-XL 10r1.1))Type "help" for help.postgres= \c lei;You are now connected to database "lei" as user "postgres".lei= select from test1;id | name----+------3 | 测试4 | 测试7 | 测试(3 rows) 两个节点的数据加起来整个8条，没有问题。 至此Postgre-XL集群搭建完成。 创建数据库、表时可能会出现以下错误： ERROR: Failed to get pooled connections 是因为pg_hba.conf配置不对，所有节点加上host all all 192.168.20.0/0 trust并重启集群即可。 ERROR: No Datanode defined in cluster 首先确认是否创建了数据节点，也就是create node相关的命令。如果创建了则执行select pgxc_pool_reload();使其生效即可。 集群管理与应用 表类型说明 REPLICATION表：各个datanode节点中，表的数据完全相同，也就是说，插入数据时，会分别在每个datanode节点插入相同数据。读数据时，只需要读任意一个datanode节点上的数据。 建表语法： CREATE TABLE repltab (col1 int, col2 int) DISTRIBUTE BY REPLICATION; DISTRIBUTE ：会将插入的数据，按照拆分规则，分配到不同的datanode节点中存储，也就是sharding技术。每个datanode节点只保存了部分数据，通过coordinate节点可以查询完整的数据视图。 CREATE TABLE disttab(col1 int, col2 int, col3 text) DISTRIBUTE BY HASH(col1); 模拟数据插入 任意登录一个coordinate节点进行建表操作 [postgres@gtm ~]$ psql -h xl1 -p 5432 -U postgrespostgres= INSERT INTO disttab SELECT generate_series(1,100), generate_series(101, 200), 'foo';INSERT 0 100postgres= INSERT INTO repltab SELECT generate_series(1,100), generate_series(101, 200);INSERT 0 100 查看数据分布结果： DISTRIBUTE表分布结果 postgres= SELECT xc_node_id, count() FROM disttab GROUP BY xc_node_id;xc_node_id | count ------------+-------1148549230 | 42-927910690 | 58(2 rows) REPLICATION表分布结果 postgres= SELECT count() FROM repltab;count -------100(1 row) 查看另一个datanode2中repltab表结果 [postgres@datanode2 pgxl9.5]$ psql -p 15432psql (PGXL 10r1.1, based on PG 10.6 (Postgres-XL 10r1.1))Type "help" for help.postgres= SELECT count() FROM repltab;count -------100(1 row) 结论：REPLICATION表中，datanode1,datanode2中表是全部数据，一模一样。而DISTRIBUTE表，数据散落近乎平均分配到了datanode1,datanode2节点中。 新增数据节点与数据重分布 在线新增节点、并重新分布数据。 新增datanode节点 在gtm集群管理节点上执行pgxc_ctl命令 [postgres@gtm ~]$ pgxc_ctl/bin/bashInstalling pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Installing pgxc_ctl_bash script as /home/postgres/pgxc_ctl/pgxc_ctl_bash.Reading configuration using /home/postgres/pgxc_ctl/pgxc_ctl_bash --home /home/postgres/pgxc_ctl --configuration /home/postgres/pgxc_ctl/pgxc_ctl.confFinished reading configuration. PGXC_CTL START Current directory: /home/postgres/pgxc_ctlPGXC 在服务器xl3上，新增一个master角色的datanode节点，名称是datanode3 端口号暂定5430，pool master暂定6669 ，指定好数据目录位置，从两个节点升级到3个节点，之后要写3个none none应该是datanodeSpecificExtraConfig或者datanodeSpecificExtraPgHba配置PGXC add datanode master datanode3 xl3 15432 6671 /home/postgres/pgxc/nodes/datanode/datanode3 none none none 等待新增完成后，查询集群节点状态： postgres= select from pgxc_node;node_name | node_type | node_port | node_host | nodeis_primary | nodeis_preferred | node_id-----------+-----------+-----------+-----------+----------------+------------------+-------------datanode1 | D | 15432 | xl1 | t | f | 888802358datanode2 | D | 15432 | xl2 | f | f | -905831925datanode3 | D | 15432 | xl3 | f | f | -705831925coord1 | C | 5432 | xl1 | f | f | 1885696643coord2 | C | 5432 | xl2 | f | f | -1197102633(4 rows) 节点新增完毕 数据重新分布 由于新增节点后无法自动完成数据重新分布，需要手动操作。 DISTRIBUTE表分布在了node1,node2节点上，如下： postgres= SELECT xc_node_id, count() FROM disttab GROUP BY xc_node_id;xc_node_id | count ------------+-------1148549230 | 42-927910690 | 58(2 rows) 新增一个节点后，将sharding表数据重新分配到三个节点上，将repl表复制到新节点 重分布sharding表postgres= ALTER TABLE disttab ADD NODE (datanode3);ALTER TABLE 复制数据到新节点postgres= ALTER TABLE repltab ADD NODE (datanode3);ALTER TABLE 查看新的数据分布： postgres= SELECT xc_node_id, count() FROM disttab GROUP BY xc_node_id;xc_node_id | count ------------+--------700122826 | 36-927910690 | 321148549230 | 32(3 rows) 登录datanode3(新增的时候，放在了xl3服务器上，端口15432)节点查看数据： [postgres@gtm ~]$ psql -h xl3 -p 15432 -U postgrespsql (PGXL 10r1.1, based on PG 10.6 (Postgres-XL 10r1.1))Type "help" for help.postgres= select count() from repltab;count -------100(1 row) 很明显,通过 ALTER TABLE tt ADD NODE (dn)命令，可以将DISTRIBUTE表数据重新分布到新节点，重分布过程中会中断所有事务。可以将REPLICATION表数据复制到新节点。 从datanode节点中回收数据 postgres= ALTER TABLE disttab DELETE NODE (datanode3);ALTER TABLEpostgres= ALTER TABLE repltab DELETE NODE (datanode3);ALTER TABLE 删除数据节点 Postgresql-XL并没有检查将被删除的datanode节点是否有replicated/distributed表的数据，为了数据安全，在删除之前需要检查下被删除节点上的数据，有数据的话，要回收掉分配到其他节点，然后才能安全删除。删除数据节点分为四步骤： 1.查询要删除节点dn3的oid postgres= SELECT oid, FROM pgxc_node;oid | node_name | node_type | node_port | node_host | nodeis_primary | nodeis_preferred | node_id -------+-----------+-----------+-----------+-----------+----------------+------------------+-------------11819 | coord1 | C | 5432 | datanode1 | f | f | 188569664316384 | coord2 | C | 5432 | datanode2 | f | f | -119710263316385 | node1 | D | 5433 | datanode1 | f | t | 114854923016386 | node2 | D | 5433 | datanode2 | f | f | -92791069016397 | dn3 | D | 5430 | datanode1 | f | f | -700122826(5 rows) 2.查询dn3对应的oid中是否有数据 testdb= SELECT FROM pgxc_class WHERE nodeoids::integer[] @> ARRAY[16397];pcrelid | pclocatortype | pcattnum | pchashalgorithm | pchashbuckets | nodeoids ---------+---------------+----------+-----------------+---------------+-------------------16388 | H | 1 | 1 | 4096 | 16397 16385 1638616394 | R | 0 | 0 | 0 | 16397 16385 16386(2 rows) 3.有数据的先回收数据 postgres= ALTER TABLE disttab DELETE NODE (dn3);ALTER TABLEpostgres= ALTER TABLE repltab DELETE NODE (dn3);ALTER TABLEpostgres= SELECT FROM pgxc_class WHERE nodeoids::integer[] @> ARRAY[16397];pcrelid | pclocatortype | pcattnum | pchashalgorithm | pchashbuckets | nodeoids ---------+---------------+----------+-----------------+---------------+----------(0 rows) 4.安全删除dn3 PGXC$ remove datanode master dn3 clean 故障节点FAILOVER 1.查看当前集群状态 [postgres@gtm ~]$ psql -h xl1 -p 5432psql (PGXL 10r1.1, based on PG 10.6 (Postgres-XL 10r1.1))Type "help" for help.postgres= SELECT oid, FROM pgxc_node;oid | node_name | node_type | node_port | node_host | nodeis_primary | nodeis_preferred | node_id-------+-----------+-----------+-----------+-----------+----------------+------------------+-------------11739 | coord1 | C | 5432 | xl1 | f | f | 188569664316384 | coord2 | C | 5432 | xl2 | f | f | -119710263316387 | datanode2 | D | 15432 | xl2 | f | f | -90583192516388 | datanode1 | D | 15432 | xl1 | t | t | 888802358(4 rows) 2.模拟datanode1节点故障 直接关闭即可 PGXC stop -m immediate datanode master datanode1Stopping datanode master datanode1.Done. 3.测试查询 只要查询涉及到datanode1上的数据，那么该查询就会报错 postgres= SELECT xc_node_id, count() FROM disttab GROUP BY xc_node_id;WARNING: failed to receive file descriptors for connectionsERROR: Failed to get pooled connectionsHINT: This may happen because one or more nodes are currently unreachable, either because of node or network failure.Its also possible that the target node may have hit the connection limit or the pooler is configured with low connections.Please check if all nodes are running fine and also review max_connections and max_pool_size configuration parameterspostgres= SELECT xc_node_id, FROM disttab WHERE col1 = 3;xc_node_id | col1 | col2 | col3------------+------+------+-------905831925 | 3 | 103 | foo(1 row) 测试发现，查询范围如果涉及到故障的node1节点，会报错，而查询的数据范围不在node1上的话，仍然可以查询。 4.手动切换 要想切换，必须要提前配置slave节点。 PGXC$ failover datanode node1 切换完成后，查询集群 postgres= SELECT oid, FROM pgxc_node;oid | node_name | node_type | node_port | node_host | nodeis_primary | nodeis_preferred | node_id -------+-----------+-----------+-----------+-----------+----------------+------------------+-------------11819 | coord1 | C | 5432 | datanode1 | f | f | 188569664316384 | coord2 | C | 5432 | datanode2 | f | f | -119710263316386 | node2 | D | 15432 | datanode2 | f | f | -92791069016385 | node1 | D | 15433 | datanode2 | f | t | 1148549230(4 rows) 发现datanode1节点的ip和端口都已经替换为配置的slave了。 本篇文章为转载内容。原文链接：https://blog.csdn.net/qianglei6077/article/details/94379331。 该文由互联网用户投稿提供，文中观点代表作者本人意见，并不代表本站的立场。 作为信息平台，本站仅提供文章转载服务，并不拥有其所有权，也不对文章内容的真实性、准确性和合法性承担责任。 如发现本文存在侵权、违法、违规或事实不符的情况，请及时联系我们，我们将第一时间进行核实并删除相应内容。

... (Unified Shared Memory USM) USM语法 数据依赖 wait() depends_on in_order queue property 练习1：事件依赖 练习2：事件依赖 UMS实验 oneAPI编程模型 oneAPI编程模型提供了一个全面、统一的开发人员工具组合，可用于各种硬件设备，其中包括跨多个工作负载领域的一系列性能库。这些库包括面向各目标架构而定制化代码的函数，因此相同的函数调用可为各种支持的架构提供优化的性能。DPC++基于行业标准和开放规范，旨在鼓励生态系统的协作和创新。 多架构编程面临的挑战 在以数据为中心的环境中，专用工作负载的数量不断增长。专用负载通常因为没有通用的编程语言或API而需要使用不同的语言和库进行编程，这就需要维护各自独立的代码库。 由于跨平台的工具支持不一致，因此开发人员必须学习和使用一整套不同的工具。单独投入精力给每种硬件平台开发软件。 oneAPI则可以利用一种统一的编程模型以及支持并行性的库，支持包括CPU、GPU、FPGA等硬件等同于原生高级语言的开发性能，并且可以与现有的HPC编程模型交互。 SYCL SYCL支持C++数据并行编程，SYCL和OpenCL一样都是由Khronos Group管理的，SYCL是建立在OpenCL之上的跨平台抽象层，支持用C++用单源语言方式编写用于异构处理器的与设备无关的代码。 DPC++ DPC++(Data Parallel C++)是一种单源语言，可以将主机代码和异构加速器内核写在同一个文件当中，在主机中调用DPC++程序，计算由加速器执行。DPC++代码简洁且效率高，并且是开源的。现有的CUDA应用、Fortran应用、OpenCL应用都可以用不同方式很方便地迁移到DPC++当中。 下图显示了原来使用不同架构的HPC开发人员的一些推荐的转换方法。 编译和运行DPC++程序 编译和运行DPC++程序主要包括三步： 初始化环境变量 编译DPC++源代码 运行程序 例如本地运行，在本地系统上安装英特尔基础工具套件，使用以下命令编译和运行DPC++程序。 source /opt/intel/inteloneapi/setvars.shdpcpp simple.cpp -o simple./simple 编程实例 实现矢量加法 以下实例描述了使用DPC++实现矢量加法的过程和源代码。 queue类 queue类用来提交给SYCL执行的命令组，是将作业提交到运算设备的一种机制，多个queue可以映射到同一个设备。 Parallel kernel Parallel kernel允许代码并行执行，对于一个不具有相关性的循环数据操作，可以用Parallel kernel并行实现 在C++代码中的循环实现 for(int i=0; i < 1024; i++){a[i] = b[i] + c[i];}); 在Parallel kernel中的并行实现 h.parallel_for(range<1>(1024), [=](id<1> i){A[i] = B[i] + C[i];}); 通用的并行编程模板 h.parallel_for(range<1>(1024), [=](id<1> i){// CODE THAT RUNS ON DEVICE }); range用来生成一个迭代序列，1为步长，在循环体中，i表示索引。 Host Accessor Host Accessor是使用主机缓冲区访问目标的访问器，它使访问的数据可以在主机上使用。通过构建Host Accessor可以将数据同步回主机，除此之外还可以通过销毁缓冲区将数据同步回主机。 buf是存储数据的缓冲区。 host_accessor b(buf,read_only); 除此之外还可以将buf设置为局部变量，当系统超出buf生存期，buf被销毁，数据也将转移到主机中。 矢量相加源代码 根据上面的知识，这里展示了利用DPC++实现矢量相加的代码。 //第一行在jupyter中指明了该cpp文件的保存位置%%writefile lab/vector_add.cppinclude <CL/sycl.hpp>using namespace sycl;int main() {const int N = 256;// 初始化两个队列并打印std::vector<int> vector1(N, 10);std::cout<<"\nInput Vector1: "; for (int i = 0; i < N; i++) std::cout << vector1[i] << " ";std::vector<int> vector2(N, 20);std::cout<<"\nInput Vector2: "; for (int i = 0; i < N; i++) std::cout << vector2[i] << " ";// 创建缓存区buffer vector1_buffer(vector1);buffer vector2_buffer(vector2);// 提交矢量相加任务queue q;q.submit([&](handler &h) {// 为缓存区创建访问器accessor vector1_accessor (vector1_buffer,h);accessor vector2_accessor (vector2_buffer,h);h.parallel_for(range<1>(N), [=](id<1> index) {vector1_accessor[index] += vector2_accessor[index];});});// 创建主机访问器将设备中数据拷贝到主机当中host_accessor h_a(vector1_buffer,read_only);std::cout<<"\nOutput Values: ";for (int i = 0; i < N; i++) std::cout<< vector1[i] << " ";std::cout<<"\n";return 0;} 运行结果 统一共享内存 (Unified Shared Memory USM) 统一共享内存是一种基于指针的方法，是将CPU内存和GPU内存进行统一的虚拟化方法，对于C++来说，指针操作内存是很常规的方式，USM也可以最大限度的减少C++移植到DPC++的代价。 下图显示了非USM(左)和USM(右)的程序员开发视角。 类型 函数调用 说明 在主机上可访问 在设备上可访问 设备 malloc_device 在设备上分配（显式） 否 是 主机 malloc_host 在主机上分配（隐式） 是 是 共享 malloc_shared 分配可以在主机和设备之间迁移（隐式） 是 是 USM语法 初始化： int data = malloc_shared<int>(N, q); int data = static_cast<int >(malloc_shared(N sizeof(int), q)); 释放 free(data,q); 使用共享内存之后，程序将自动在主机和运算设备之间隐式移动数据。 数据依赖 使用USM时，要注意数据之间的依赖关系以及事件之间的依赖关系，如果两个线程同时修改同一个内存区，将产生不可预测的结果。 我们可以使用不同的选项管理数据依赖关系： 内核任务中的 wait() 使用 depends_on 方法 使用 in_queue 队列属性 wait() q.submit([&](handler &h) {h.parallel_for(range<1>(N), [=](id<1> i) { data[i] += 2; });}).wait(); // <--- wait() will make sure that task is complete before continuingq.submit([&](handler &h) {h.parallel_for(range<1>(N), [=](id<1> i) { data[i] += 3; });}); depends_on auto e = q.submit([&](handler &h) { // <--- e is event for kernel taskh.parallel_for(range<1>(N), [=](id<1> i) { data[i] += 2; });});q.submit([&](handler &h) {h.depends_on(e); // <--- waits until event e is completeh.parallel_for(range<1>(N), [=](id<1> i) { data[i] += 3; });}); in_order queue property queue q(property_list{property::queue::in_order()}); // <--- this will make sure all the task with q are executed sequentially 练习1：事件依赖 以下代码使用 USM，并有三个提交到设备的内核。每个内核修改相同的数据阵列。三个队列之间没有数据依赖关系 为每个队列提交添加 wait() 在第二个和第三个内核任务中实施 depends_on() 方法 使用 in_order 队列属性，而非常规队列： queue q{property::queue::in_order()}; %%writefile lab/usm_data.cppinclude <CL/sycl.hpp>using namespace sycl;static const int N = 256;int main() {queue q{property::queue::in_order()};//用队列限制执行顺序std::cout << "Device : " << q.get_device().get_info<info::device::name>() << "\n";int data = static_cast<int >(malloc_shared(N sizeof(int), q));for (int i = 0; i < N; i++) data[i] = 10;q.parallel_for(range<1>(N), [=](id<1> i) { data[i] += 2; });q.parallel_for(range<1>(N), [=](id<1> i) { data[i] += 3; });q.parallel_for(range<1>(N), [=](id<1> i) { data[i] += 5; });q.wait();//wait阻塞进程for (int i = 0; i < N; i++) std::cout << data[i] << " ";std::cout << "\n";free(data, q);return 0;} 执行结果 练习2：事件依赖 以下代码使用 USM，并有三个提交到设备的内核。前两个内核修改了两个不同的内存对象，第三个内核对前两个内核具有依赖性。三个队列之间没有数据依赖关系 %%writefile lab/usm_data2.cppinclude <CL/sycl.hpp>using namespace sycl;static const int N = 1024;int main() {queue q;std::cout << "Device : " << q.get_device().get_info<info::device::name>() << "\n";//设备选择int data1 = malloc_shared<int>(N, q);int data2 = malloc_shared<int>(N, q);for (int i = 0; i < N; i++) {data1[i] = 10;data2[i] = 10;}auto e1 = q.parallel_for(range<1>(N), [=](id<1> i) { data1[i] += 2; });auto e2 = q.parallel_for(range<1>(N), [=](id<1> i) { data2[i] += 3; });//e1,e2指向两个事件内核q.parallel_for(range<1>(N),{e1,e2}, [=](id<1> i) { data1[i] += data2[i]; }).wait();//depend on e1,e2for (int i = 0; i < N; i++) std::cout << data1[i] << " ";std::cout << "\n";free(data1, q);free(data2, q);return 0;} 运行结果 UMS实验 在主机中初始化两个vector，初始数据为25和49，在设备中初始化两个vector，将主机中的数据拷贝到设备当中，在设备当中并行计算原始数据的根号值，然后将data1_device和data2_device的数值相加，最后将数据拷贝回主机当中，检验最后相加的和是否是12，程序结束前将内存释放。 %%writefile lab/usm_lab.cppinclude <CL/sycl.hpp>include <cmath>using namespace sycl;static const int N = 1024;int main() {queue q;std::cout << "Device : " << q.get_device().get_info<info::device::name>() << "\n";//intialize 2 arrays on hostint data1 = static_cast<int >(malloc(N sizeof(int)));int data2 = static_cast<int >(malloc(N sizeof(int)));for (int i = 0; i < N; i++) {data1[i] = 25;data2[i] = 49;}// STEP 1 : Create USM device allocation for data1 and data2int data1_device = static_cast<int >(malloc_device(N sizeof(int),q));int data2_device = static_cast<int >(malloc_device(N sizeof(int),q));// STEP 2 : Copy data1 and data2 to USM device allocationq.memcpy(data1_device, data1, sizeof(int) N).wait();q.memcpy(data2_device, data2, sizeof(int) N).wait();// STEP 3 : Write kernel code to update data1 on device with sqrt of valueauto e1 = q.parallel_for(range<1>(N), [=](id<1> i) { data1_device[i] = std::sqrt(25); });auto e2 = q.parallel_for(range<1>(N), [=](id<1> i) { data2_device[i] = std::sqrt(49); });// STEP 5 : Write kernel code to add data2 on device to data1q.parallel_for(range<1>(N),{e1,e2}, [=](id<1> i) { data1_device[i] += data2_device[i]; }).wait();// STEP 6 : Copy data1 on device to hostq.memcpy(data1, data1_device, sizeof(int) N).wait();q.memcpy(data2, data2_device, sizeof(int) N).wait();// verify resultsint fail = 0;for (int i = 0; i < N; i++) if(data1[i] != 12) {fail = 1; break;}if(fail == 1) std::cout << " FAIL"; else std::cout << " PASS";std::cout << "\n";// STEP 7 : Free USM device allocationsfree(data1_device, q);free(data1);free(data2_device, q);free(data2);// STEP 8 : Add event based kernel dependency for the Steps 2 - 6return 0;} 运行结果 本篇文章为转载内容。原文链接：https://blog.csdn.net/MCKZX/article/details/127630566。 该文由互联网用户投稿提供，文中观点代表作者本人意见，并不代表本站的立场。 作为信息平台，本站仅提供文章转载服务，并不拥有其所有权，也不对文章内容的真实性、准确性和合法性承担责任。 如发现本文存在侵权、违法、违规或事实不符的情况，请及时联系我们，我们将第一时间进行核实并删除相应内容。

本文详细探讨了Netty框架如何利用JIT编译器优化性能，重点介绍了ChannelPipeline和ByteBuf的设计。通过内联技术和逃逸分析，Netty减少了函数调用开销和垃圾回收压力，提升了整体性能。文章还强调了ByteBuf在内存管理中的作用，并指出JIT编译器能够有效优化频繁调用的方法，从而实现高效的堆内存使用。

本文详细解析了Netty在故障恢复方面的实现机制，包括异常处理与重试机制、零拷贝技术优化内存管理、长连接复用结合心跳检测保障连接稳定性，以及数据缓冲与批量处理提升效率。Netty通过这些关键技术，有效应对网络异常，确保系统高性能与高可靠性，为分布式系统开发提供了坚实支撑。

本文详细介绍了Netty框架在大数据流处理中的应用，强调其异步非阻塞I/O和零拷贝技术能大幅提升高吞吐量和低延迟。通过灵活的消息编解码与容错重试机制，Netty有效应对大数据场景下的复杂需求，展现其扩展性和高性能网络编程的优势。文中结合实例分析，指出Netty虽有学习门槛，但其核心能力对大数据流处理至关重要。

这篇文章详细介绍了如何在网站或博客中嵌入优酷视频的HTML代码。首先，通过点击优酷视频页面的“分享”按钮复制HTML代码。然后将复制的代码粘贴到iframe标签中，并调整height和width属性以适应网页布局。其中，src属性应替换为实际要嵌入的视频链接。此方法同样适用于其他视频网站的视频嵌入操作，旨在帮助用户便捷地在个人网站/博客中展示优酷等平台的视频内容。

Netty作为高性能Java NIO框架，其ByteBuf内存管理机制在设计上独具匠心。通过HeapByteBuf和DirectByteBuf两种类型实现，灵活分配堆内和堆外内存；采用内存池(PooledByteBufAllocator)技术高效复用内存块以降低GC压力；智能扩容策略减少数据拷贝次数；同时运用volatile变量与AtomicIntegerFieldUpdater控制内存碎片，优化并发环境下ByteBuf操作的安全性和内存利用率。深入理解并合理应用这些机制，对于构建稳定高效的网络应用服务具有重要意义。

...和Protocol Buffers（protobuf）数据格式。gRPC支持多种语言环境，并提供了客户端与服务器之间直接高效的双向流式通信功能。在本文语境中，gRPC集成到RabbitMQ意味着可以通过适配器或桥接器将gRPC请求转化为RabbitMQ可识别的消息进行处理，实现服务间的异步通信。 HTTP API Gateway , HTTP API Gateway是一种架构模式，它充当了系统的入口点，集中处理来自客户端的所有HTTP请求，并负责转发、转换这些请求到相应的后端服务。在RabbitMQ与HTTP集成的场景中，API Gateway接收客户端的HTTP请求，然后将这些请求封装成RabbitMQ可以理解的消息格式，发布到特定的交换机，从而实现在分布式系统中的服务解耦和异步处理。

share.js是一个专注于简化网页内容分享过程的轻量级jQuery插件。它旨在提供一种美观且易于集成的解决方案，以增强用户在浏览网页时的互动体验。通过引入socialsharing和tooltip功能，该插件不仅使分享操作变得更加便捷，还显著提升了用户体验。功能亮点1.广泛兼容性：支持所有主流社交媒体平台，包括但不限于Facebook、Twitter、LinkedIn、Pinterest等，确保用户可以轻松选择最符合其需求的分享渠道。2.美观设计：share.js采用简洁而现代的设计风格，与各种网站风格相融合，不会破坏原有网页的整体美观性。其分享按钮既实用又不失时尚感，提升了用户在分享过程中的视觉体验。3.交互式提示：插件内置了智能提示功能，当用户悬停在按钮上时，会显示具体的分享信息或相关链接，增强了用户的理解与操作效率。4.灵活性与定制化：开发者可根据需要调整样式、布局和行为，以适应不同的网站需求和设计偏好。这使得share.js能够无缝融入各类网站环境，提供个性化的分享体验。5.轻量级代码：作为一款轻量级插件，share.js对页面性能的影响极小，加载速度快，兼容各种浏览器，保证了良好的用户体验和网站性能。6.易于集成：通过简单的JavaScript调用即可快速将分享按钮添加至网页中，无需复杂的配置或额外的依赖库，极大地降低了开发成本和时间。应用场景-博客与新闻网站：促进内容的广泛传播，增加文章的可见度和读者互动。-电子商务平台：鼓励用户分享商品信息，提高产品曝光率和转化率。-社交媒体整合：在个人主页或社区板块中，方便用户分享动态、文章或图片，增强社区活跃度。总之，share.js通过其强大的功能集、美观的界面设计和简便的集成方式，成为了一款不可或缺的工具，对于任何希望提升用户参与度和内容分享效率的网站开发者来说，都是一款值得尝试的优秀插件。

Python编程语言在处理中文字符时，由于默认字符集为ASCII码，因此在Python 2.x版本中需声明使用utf-8编码以正确显示中文。而在Python 3.x版本中，默认已改为utf-8编码，但仍需关注输出时的编码转换问题。面对中文字符串输出异常，可利用Unicode编码原理，通过sys.stdout.buffer.write()指定utf-8编码进行输出，或采用先转utf-8再decode为unicode_escape的简便方式来确保中文字符的准确显示。

Java IO包括BIO和NIO两种操作方式。BIO采用同步阻塞模型，当线程执行读写操作时会一直阻塞直至完成，适用于连接数少且不活跃的场景；而NIO则是异步非阻塞模式，即使IO未完成，线程也能继续执行其他任务，通过Selector管理多个通道事件，并利用SocketChannel进行非阻塞读写，提升了系统资源利用率和并发处理能力，适合于连接数多且相对不活跃的情况，如长连接、心跳检测等场景。

本文针对Netty服务器在遭遇网络中断问题时，提出了具体的解决方案。首先阐述了Netty基于NIO实现的异步事件驱动模型，当网络发生异常如ChannelException或UnresolvedAddressException时，通过监听ChannelFuture的状态变化（利用FutureListener），可以及时感知并处理连接失败的情况。同时，引入心跳检测机制，客户端定期发送心跳包以检测网络是否畅通；若中断，则启动重连机制，设定重试次数与间隔时间，确保快速重新建立连接。这些策略共同保障了Netty服务器在网络不稳定情况下的正常运行和高效恢复。

本文介绍了在Windows环境中使用PowerShell对文件系统审核日志进行高效筛选的方法。针对日志量大、查询不便的问题，通过启用文件系统审核功能，并配置日志轮替归档策略（如500MB大小、保留60天）。运用Get-WinEvent命令结合FilterXPath参数，可精准筛选特定事件ID（如4660）及文件删除操作（AccessMask为0x10000）的日志记录，甚至细化到指定用户的操作。此外，还提供了从保存的事件日志文件中筛选以及按时间范围筛选的安全性日志方法，并给出了脚本示例以自动管理过期存档日志。同时，文章还列出了文件操作码表便于理解和解析日志内容。

本文针对HBase性能测试与调优方法进行了深入探讨，首先从基准测试工具（如YCSB）入手，模拟极端场景对读写吞吐量、响应时间及并发处理能力进行评估。在实际调优过程中，重点关注RegionServer配置优化，包括网络参数调整、内存资源分配（如BlockCache和MemStore）以及动态调整Region大小以缓解热点问题。同时，架构层面可通过增加RegionServer节点和预分裂策略提升系统性能。此外，强调了数据模型优化的重要性，如合理设计RowKey分散热点，选择合适的列族压缩算法以降低存储占用。通过实践案例分析，展示了如何结合业务特征精准定位问题并采取有效调优措施，从而显著提升HBase服务的性能表现。

在.NET 2.0时代，大容量文件上传问题通过修改web.config或使用Flash等方法解决，但存在操作复杂、稳定性差等问题。进入HTML5时代后，借助Plupload这一前端脚本库，利用分片上传技术有效简化了该问题。Plupload能自动检测浏览器支持情况，并依次采用HTML5、Flash或Silverlight进行文件上传，同时提供丰富的事件监听功能，如文件选取、上传进度、成功及失败回调等。开发者可通过Install-Package Plupload便捷安装，并结合UploadCoursePackage.ashx处理分片与不分片的上传逻辑，实现安全高效的文件上传及进度展示。

本文详细介绍了在SpringBoot项目中利用HttpServletResponse和HttpServletRequest实现文件切片下载功能，通过处理Range请求头实现断点续传。后端Java代码展示了如何根据用户输入的文件名读取文件流，并使用InputStream、OutputStream进行文件分片传输。前端部分采用Vue框架展示交互界面，用户可以暂停与开始下载操作，以提升大文件下载体验。关键词：SpringBoot、HttpServletResponse、HttpServletRequest、InputStream、OutputStream、文件切片下载、分片传输、断点续传、Range请求头、Vue前端。

本文聚焦Android应用中的卡顿优化问题，从用户感知的界面不流畅到技术层面的UI绘制流程、VSYNC信号同步机制，深入剖析了卡顿产生的根源。涉及关键词包括：卡顿优化、Android应用、UI绘制、16ms帧率、VSYNC信号、SurfaceFlinger、Choreographer等。文中强调了在16ms内完成渲染以保证流畅体验，并详述了如何通过监控Choreographer计算掉帧情况以识别卡顿原因，同时关注CPU使用率与GPU处理效率对性能的影响，提倡动画优化和系统资源管理来降低卡顿现象，提升用户体验。

开源软件生态中，普遍存在组织作为索取者而非贡献者的现象，加剧了维护者负担。以Log4j为例，尽管广泛应用且下载量巨大，但主要依赖少数核心人员维护，凸显出开源项目对志愿者过度依赖的问题，即“公地悲剧”。为解决这一问题，建立内部开源计划办公室（OSPO）显得尤为重要，它有助于关键开源项目的健康发展、降低技术负债并增强企业招聘吸引力。通过参与开源社区并设立OSPO，企业既能确保项目安全与战略一致性，也能从开源活动中获得商业优势和战略伙伴关系，从而实现开源项目健康、可持续发展。

该文章介绍了如何在jQuery的$.ajax方法中设置HTTP请求头。首先，通过直接在Ajax选项的headers对象内定义键值对来设置如Accept: application/json; charset=utf-8这样的标准请求头。其次，利用beforeSend回调函数动态添加或修改请求头，例如为sso_token设置自定义请求头值。两种方式均有助于实现客户端与服务器间的数据传输格式协商、身份验证或其他特定需求。

这篇文章详细介绍了在 MySQL 数据库中设置系统变量的方法。首先，通过 SHOW VARIABLES; 命令查看当前的系统变量值；其次，利用 SET GLOBAL VARIABLE_NAME=value; 临时修改系统变量，例如将 max_connections 设置为特定数值。若要实现重启后系统变量值仍然生效，则需编辑 MySQL 的配置文件 my.cnf，并在此文件中设定相应变量（如 max_connections），随后重启 MySQL 服务。文章提供了针对MySQL系统变量进行查看、调整与永久保存的具体步骤和命令。