学习目标
学习检查数据库运行状态、锁信息、维护表、重建索引等数据库维护操作技术
课程学习
1.检查openGauss状态
--操作系统命令查看实例状态root@modb:~# su - omm omm@modb:~$ ps -ef|grep gaussdb omm 1 0 0 14:03 ? 00:00:00 gaussdb omm 383 375 0 14:05 pts/0 00:00:00 grep gaussdbomm@modb:~$ ps -Tp 1 PID SPID TTY TIME CMD 1 1 ? 00:00:00 gaussdb 1 293 ? 00:00:00 jemalloc_bg_thd 1 299 ? 00:00:00 gaussdb 1 300 ? 00:00:00 syslogger 1 301 ? 00:00:00 jemalloc_bg_thd 1 302 ? 00:00:00 alarm 1 303 ? 00:00:00 reaper 1 304 ? 00:00:00 jemalloc_bg_thd 1 305 ? 00:00:00 jemalloc_bg_thd 1 349 ? 00:00:00 checkpointer 1 350 ? 00:00:00 Spbgwriter 1 352 ? 00:00:00 pagewriter 1 351 ? 00:00:00 pagewriter 1 353 ? 00:00:00 pagewriter 1 355 ? 00:00:00 pagewriter 1 354 ? 00:00:00 pagewriter 1 356 ? 00:00:00 WALwriter 1 357 ? 00:00:00 WALwriteraux 1 358 ? 00:00:00 AVClauncher 1 359 ? 00:00:00 Jobscheduler 1 360 ? 00:00:00 asyncundolaunch 1 361 ? 00:00:00 globalstats 1 362 ? 00:00:00 applylauncher 1 363 ? 00:00:00 statscollector 1 364 ? 00:00:00 percentworker 1 365 ? 00:00:00 ashworker 1 366 ? 00:00:00 TrackStmtWorker 1 367 ? 00:00:00 auditor 1 368 ? 00:00:00 2pccleaner 1 369 ? 00:00:00 faultmonitor 1 370 ? 00:00:00 WLMworker 1 371 ? 00:00:00 WLMmonitor 1 372 ? 00:00:00 WLMarbiter 1 373 ? 00:00:00 undorecycler--使用数据库工具gs_ctl查看实例状态 omm@modb:~$ gs_ctl status -D /var/lib/opengauss/data [2022-12-14 14:07:05.766][390][][gs_ctl]: gs_ctl status,datadir is /var/lib/opengauss/data gs_ctl: server is running (PID: 1) /usr/local/opengauss/bin/gaussdb--检查参数 omm@modb:~$ gsql -r gsql ((openGauss 3.0.0 build 02c14696) compiled at 2022-04-01 18:12:00 commit 0 last mr ) Non-SSL connection (SSL connection is recommended when requiring high-security) Type "help" for help. omm=# show data_directory; data_directory ------------------------- /var/lib/opengauss/data (1 row)
2.检查锁信息
锁机制是数据库保证数据一致性的重要手段，检查相关信息可以检查数据库的事务和运行状况。
--查询数据库中的锁信息omm=# SELECT * FROM pg_locks; locktype | database | relation | page | tuple | bucket | virtualxid | transactionid | classid | objid | objsubid | virtualtransaction | pid | sessionid | mode | granted | fastpath | locktag | global_sessionid ------------+----------+----------+------+-------+--------+------------+---------------+---------+-------+----------+--------------------+---------------- -+-----------------+-----------------+---------+----------+-------------------+------------------ relation | 16384 | 12010 | | | | | | | | | 11/56 | 281471288638704 | 281471288638704 | AccessShareLock | t | t | 4000:2eea:0:0:0:0 | 0:0#0 virtualxid | | | | | | 11/56 | | | | | 11/56 | 281471288638704 | 281471288638704 | ExclusiveLock | t | t | b:38:0:0:0:7 | 0:0#0 virtualxid | | | | | | 9/47 | | | | | 9/47 | 281471461653744 | 281471461653744 | ExclusiveLock | t | t | 9:2f:0:0:0:7 | 0:0#0 (3 rows) --查询等待锁的线程状态信息 omm=# SELECT * FROM pg_thread_wait_status WHERE wait_status = 'acquire lock'; node_name | db_name | thread_name | query_id | tid | sessionid | lwtid | psessionid | tlevel | smpid | wait_status | wait_event | locktag | lockmode | bl ock_sessionid | global_sessionid -----------+---------+-------------+----------+-----+-----------+-------+------------+--------+-------+-------------+------------+---------+----------+--- --------------+------------------ (0 rows)
3.统计事件数据
SQL语句长时间运行会占用大量系统资源，用户可以通过查看事件发生的时间，占用内存大小来了解现在数据库运行状态。
--查询事件的时间，查询事件的线程启动时间、事务启动时间、SQL启动时间以及状态变更时间。omm=# SELECT backend_start,xact_start,query_start,state_change FROM pg_stat_activity; backend_start | xact_start | query_start | state_change -------------------------------+-------------------------------+-------------------------------+------------------------------- 2022-12-14 14:07:28.892357+08 | 2022-12-14 14:09:09.657135+08 | 2022-12-14 14:09:09.657135+08 | 2022-12-14 14:09:09.65714+08 2022-12-14 14:04:16.777414+08 | | | 2022-12-14 14:04:16.777414+08 | 2022-12-14 14:04:16.780066+08 | 2022-12-14 14:04:16.780066+08 | 2022-12-14 14:04:16.780365+08 2022-12-14 14:04:16.777346+08 | | | 2022-12-14 14:04:16.771717+08 | | | 2022-12-14 14:09:09.591449+08 2022-12-14 14:04:16.771496+08 | | | 2022-12-14 14:09:07.064376+08 2022-12-14 14:04:16.771339+08 | | | 2022-12-14 14:04:16.771361+08 2022-12-14 14:04:16.770981+08 | | | 2022-12-14 14:09:09.068391+08 2022-12-14 14:04:16.765199+08 | | | (9 rows)--查询当前服务器的会话计数信息omm=# SELECT count(*) FROM pg_stat_activity; count ------- 9 (1 row)--查询系统级统计信息，查询当前使用内存最多的会话信息。omm=# SELECT * FROM pv_session_memory_detail() ORDER BY usedsize desc limit 10; sessid | threadid | contextname | level | parent | totalsize | freesize | usedsize --------+----------+-------------+-------+--------+-----------+----------+---------- (0 rows) --查询表统计信息postgres=# SELECT * FROM pg_statistic; omm=# SELECT * FROM pg_statistic; starelid | starelkind | staattnum | stainherit | stanullfrac | stawidth | stadistinct | stakind1 | stakind2 | stakind3 | stakind4 | stakind5 | staop1 | s taop2 | staop3 | staop4 | staop5 | stanumbers1 | stanumbers2 | --More--
4.数据库版本、日志信息
--数据库版本omm=# SELECT version(); version ------------------------------------------------------------------------------------------------------------------------------------------------------- (openGauss 3.0.0 build 02c14696) compiled at 2022-04-01 18:12:00 commit 0 last mr on aarch64-unknown-linux-gnu, compiled by g++ (GCC) 7.3.0, 64-bit (1 row)--在gsql中postgres=# \! ls /opt/huawei/install/app bin etc include jre lib share simpleInstall version.cfgpostgres=# \q--数据库日志位置omm=# \! ls /var/lib/opengauss/data/pg_log postgresql-2022-12-14_140405.log postgresql-2022-12-14_140416.log omm=# \q omm@modb:~$ cd /var/lib/opengauss/data/pg_log omm@modb:/var/lib/opengauss/data/pg_log$ ls postgresql-2022-12-14_140405.log postgresql-2022-12-14_140416.log
5.检查应用连接数
如果应用程序与数据库的连接数超过最大值，则新的连接无法建立。建议每天检查连接数，及时释放空闲的连接或者增加最大连接数。
--执行如下SQL语句查看连接数。omm@modb:/var/lib/opengauss/data/pg_log$ cd omm@modb:~$ gsql -r gsql ((openGauss 3.0.0 build 02c14696) compiled at 2022-04-01 18:12:00 commit 0 last mr ) Non-SSL connection (SSL connection is recommended when requiring high-security) Type "help" for help. omm=# SELECT count(*) FROM (SELECT pg_stat_get_backend_idset() AS backendid) AS s; count ------- 21 (1 row)--查看现有最大连接数。omm=# SHOW max_connections; max_connections ----------------- 200 (1 row)
6.例行维护表：VACUUM、VACUUM FULL和ANALYZE
为了保证数据库的有效运行，数据库必须在插入/删除操作后，基于客户场景，定期做VACUUM FULL和ANALYZE，更新统计信息，以便获得更优的性能。
VACUUM和ANALYZE会导致I/O流量的大幅增加，这可能会影响其他活动会话的性能。
--创建表、插入数据、查看数据omm=# CREATE TABLE customer_t omm(# ( c_customer_sk integer, omm(# c_customer_id char(5), omm-# c_first_name char(6), omm(# c_last_name char(8) omm(# ) ; CREATE TABLE omm=# INSERT INTO customer_t (c_customer_sk, c_customer_id, c_first_name,c_last_name) VALUES omm-# (6885, 1, 'Joes', 'Hunter'), omm-# (4321, 2, 'Lily','Carter'), omm-# (9527, 3, 'James', 'Cook'), omm-# (9500, 4, 'Lucy', 'Baker'); INSERT 0 4 omm=# select count(*) from customer_t; omm=# count ------- 4 (1 row)--使用VACUUM或VACUUM FULL命令，进行磁盘空间回收。 --使用VACUUMomm=# VACUUM customer_t; VACUUM--使用VACUUM full --需要向正在执行的表增加排他锁，且需要停止其他所有数据库操作。omm=# VACUUM FULL customer_t; VACUUM--使用ANALYZE语句更新统计信息。omm=# ANALYZE customer_t; ANALYZE--使用ANALYZE VERBOSE语句更新统计信息，并输出表的相关信息。 omm=# ANALYZE VERBOSE customer_t; INFO: analyzing "public.customer_t"(gaussdb pid=1) INFO: ANALYZE INFO : "customer_t": scanned 1 of 1 pages, containing 4 live rows and 0 dead rows; 4 rows in sample, 4 estimated total rows(gaussdb pid=1) ANALYZE--也可以同时执行VACUUM ANALYZE命令进行查询优化。omm=# VACUUM ANALYZE customer_t; VACUUM
维护建议：
-
定期对部分大表做VACUUM FULL，在性能下降后为全库做VACUUM FULL，目前暂定每月做一次VACUUM FULL。
-
定期对系统表做VACUUM FULL，主要是PG_ATTRIBUTE。
-
启用系统自动清理线程（AUTOVACUUM）自动执行VACUUM和ANALYZE，回收被标识为删除状态的记录空间，并更新表的统计数据。
7.重建索引
在重建索引前，用户可以通过临时增大maintenance_work_mem和psort_work_mem的取值来加快索引的重建。
--环境准备omm=# CREATE TABLESPACE EXAMPLE RELATIVE LOCATION 'tablespace1/tablespace_1'; CREATE TABLESPACE omm=# CREATE schema HR; CREATE SCHEMA omm=# CREATE TABLE HR.areaS( area_ID NUMBER, area_NAME VARCHAR2(25) )tablespace EXAMPLE; CREATE TABLE omm=# \d HR.areaS Table "hr.areas" Column | Type | Modifiers -----------+-----------------------+----------- area_id | numeric | area_name | character varying(25) | Tablespace: "example" omm=# CREATE INDEX areaS_idx ON HR.areaS (area_id); CREATE INDEX--使用REINDEX TABLE语句重建索引。omm=# REINDEX table HR.areaS; REINDEX
课程作业
1、使用操作系统命令和数据库工具检查实例状态
[omm@node1 ~]$ ps -ef|grep gaussdb omm 10482 1 45 02:56 ? 03:20:12 /opt/huawei/install/app/bin/gaussdb -D /opt/huawei/install/data/dn omm 26542 25344 0 10:12 pts/0 00:00:00 grep --color=auto gaussdb [omm@node1 ~]$ ps -Tp 1 PID SPID TTY TIME CMD 1 1 ? 00:00:02 systemd[omm@node1 ~]$ gs_ctl status -D /var/lib/opengauss/data [2022-12-14 10:12:25.344][26545][][gs_ctl]: gs_ctl status,datadir is /var/lib/opengauss/data no server running[omm@node1 ~]$ gsql -d postgres -p 26000 -r gsql ((openGauss 2.0.0 build 78689da9) compiled at 2021-03-31 21:04:03 commit 0 last mr ) Non-SSL connection (SSL connection is recommended when requiring high-security) Type "help" for help. postgres=# show data_directory; data_directory ----------------------------- /opt/huawei/install/data/dn (1 row)
2、检查锁信息
postgres=# SELECT * FROM pg_locks;
locktype | database | relation | page | tuple | bucket | virtualxid | transactionid | classid | objid | objsubid | virtualtransaction | pid | sessionid | mode | g
ranted | fastpath | locktag
------------+----------+----------+------+-------+--------+------------+---------------+---------+-------+----------+--------------------+-----------------+-----------------+-----------------+--
-------+----------+-------------------
relation | 15103 | 11835 | | | | | | | | | 6/953 | 140413633427200 | 140413633427200 | AccessShareLock | t
| t | 3aff:2e3b:0:0:0:0
virtualxid | | | | | | 6/953 | | | | | 6/953 | 140413633427200 | 140413633427200 | ExclusiveLock | t
| t | 6:3b9:0:0:0:7
(2 rows)
postgres=# SELECT * FROM pg_thread_wait_status WHERE wait_status = 'acquire lock';
node_name | db_name | thread_name | query_id | tid | sessionid | lwtid | psessionid | tlevel | smpid | wait_status | wait_event | locktag | lockmode | block_sessionid
-----------+---------+-------------+----------+-----+-----------+-------+------------+--------+-------+-------------+------------+---------+----------+-----------------
(0 rows)
3、查看数据库版本、日志信息
postgres=# SELECT version();
version
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
PostgreSQL 9.2.4 (openGauss 2.0.0 build 78689da9) compiled at 2021-03-31 21:04:03 commit 0 last mr on x86_64-unknown-linux-gnu, compiled by g++ (GCC) 7.3.0, 64-bit
(1 row)
--在gsql中
postgres=# \! ls /opt/huawei/install/app
bin etc include jre lib share simpleInstall version.cfg
postgres=# \q
[omm@node1 ~]$ cd /opt/huawei/install/app
[omm@node1 app]$ ls
bin etc include jre lib share simpleInstall version.cfg
4、检查应用连接数、查看现有最大连接数
[omm@node1 ~]$ gsql -d postgres -p 26000 -r
gsql ((openGauss 2.0.0 build 78689da9) compiled at 2021-03-31 21:04:03 commit 0 last mr )
Non-SSL connection (SSL connection is recommended when requiring high-security)
Type "help" for help.
postgres=# SELECT count(*) FROM (SELECT pg_stat_get_backend_idset() AS backendid) AS s;
count
-------
15
(1 row)
postgres=# SHOW max_connections;
max_connections
-----------------
5000
(1 row)
5、对表进行VACUUM、VACUUM FULL和ANALYZE操作
postgres=# CREATE TABLE customer_t
postgres-# ( c_customer_sk integer,
postgres(# c_customer_id char(5),
postgres(# c_first_name char(6),
postgres(# c_last_name char(8)
postgres(# ) ;
CREATE TABLE
postgres=# INSERT INTO customer_t (c_customer_sk, c_customer_id, c_first_name,c_last_name) VALUES
postgres-# (6885, 1, 'Joes', 'Hunter'),
postgres-# (4321, 2, 'Lily','Carter'),
postgres-# (9527, 3, 'James', 'Cook'),
postgres-# (9500, 4, 'Lucy', 'Baker');
INSERT 0 4
postgres=# select count(*) from customer_t;
count
-------
4
(1 row)
postgres=# VACUUM customer_t;
VACUUM
postgres=# VACUUM FULL customer_t;
VACUUM
postgres=# ANALYZE customer_t;
ANALYZE
postgres=# ANALYZE VERBOSE customer_t;
INFO: analyzing "public.customer_t"(dn_6001 pid=10482)
INFO: ANALYZE INFO : "customer_t": scanned 1 of 1 pages, containing 4 live rows and 0 dead rows; 4 rows in sample, 4 estimated total rows(dn_6001 pid=10482)
ANALYZE
postgres=# VACUUM ANALYZE customer_t;
VACUUM
6、建表、索引，进行重建索引操作
postgres=# CREATE TABLESPACE EXAMPLE RELATIVE LOCATION 'tablespace1/tablespace_1';
CREATE TABLESPACE
postgres=# CREATE schema HR;
CREATE SCHEMA
postgres=# CREATE TABLE HR.areaS( area_ID NUMBER, area_NAME VARCHAR2(25) )tablespace EXAMPLE;
CREATE TABLE
postgres=# \d HR.areaS
Table "hr.areas"
Column | Type | Modifiers
-----------+-----------------------+-----------
area_id | numeric |
area_name | character varying(25) |
Tablespace: "example"
postgres=# CREATE INDEX areaS_idx ON HR.areaS (area_id);
CREATE INDEX
postgres=# REINDEX table HR.areaS;
REINDEX
最后修改时间：2022-12-14 14:18:05
「喜欢这篇文章，您的关注和赞赏是给作者最好的鼓励」
关注作者
【版权声明】本文为墨天轮用户原创内容，转载时必须标注文章的来源（墨天轮），文章链接，文章作者等基本信息，否则作者和墨天轮有权追究责任。如果您发现墨天轮中有涉嫌抄袭或者侵权的内容，欢迎发送邮件至：contact@modb.pro进行举报，并提供相关证据，一经查实，墨天轮将立刻删除相关内容。
