介绍一下使用pg_dump与pg_restore来迁移数据时的基本步骤。
1. 首先当然是准备目标环境,修改数据库参数。
2. 更改目标环境的配置（如
archive_mode = off
log_min_duration_statement = -1
log_statement = 'none'
checkpoint_segments = 256
checkpoint_timeout = 1h
checkpoint_completion_target = 0.1
autovacuum = off)
原则是尽量减少导入数据的开销,加快导入速度.
2. 停止业务，或者理解为停止对源库的任何更改操作，保持数据一致。（这里有个特例，如果数据只有插入，并且有PK，PK不会被更改与删除时，也可以继续此类业务，迁移完将增量数据补到目标库即可）
3. 备份源数据库
4. 还原目标数据
5. 修正异常
6. analyze目标数据库
7. 修改参数，重启目标数据。
前面的那几个参数改成OLTP适应的参数。
8. 停止源库，切换业务到目标数据库
9.（如果有增量则同步增量，同步完analyze目标数据库）
10. 调整表空间.
以下是一个普通的配置
listen_addresses = '*'
port = 1921
max_connections = 2400
superuser_reserved_connections = 13
unix_socket_permissions = 0700
password_encryption = on
shared_buffers = 2048MB
maintenance_work_mem = 2048MB
max_stack_depth = 8MB
max_files_per_process = 10000
vacuum_cost_delay = 10ms
effective_io_concurrency = 6
wal_sync_method = open_sync
wal_buffers = 16384kB
checkpoint_segments = 32
checkpoint_timeout = 5min
checkpoint_completion_target = 0.5
random_page_cost = 2.0
effective_cache_size = 35000MB
default_statistics_target = 300
constraint_exclusion = partition
logging_collector = on
log_directory = '/var/applog/pg_log'
log_filename = 'postgresql-%Y-%m-%d_%H%M%S.log'
log_truncate_on_rotation = on
log_rotation_age = 1d
log_rotation_size = 50MB
log_min_duration_statement = 1000ms
log_checkpoints = on
log_line_prefix = '%t %u@%d (PID:%p SID:%c) %r '
log_lock_waits = on
log_statement = 'ddl'
stats_temp_directory = '/database/pgdata/digoal/pg_stat_tmp'
autovacuum = on
log_autovacuum_min_duration = 10000
autovacuum_max_workers = 13
autovacuum_naptime = 1min
default_tablespace = 'tbs_default'
temp_tablespaces = 'tbs_temp'
datestyle = 'iso, mdy'
lc_messages = 'C'
lc_monetary = 'C'
lc_numeric = 'C'
lc_time = 'C'
default_text_search_config = 'pg_catalog.english'
deadlock_timeout = 1s
