当我更新特别大的表时，更新超时，因为在索引重建时表被锁定。有没有办法重建索引ONLINE（即Oracle），以便更新不会超时？
答案 0 :(得分：0)
简单回答：不，没有办法。
更复杂的答案：您可以通过使用基于语句的复制并首先将索引添加到从属服务器，然后将其作为主服务器来模拟在线索引添加。这就是人们使用http://mysql-mmm.org/等软件包的原因。
答案 1 :(得分：0)
{
"caption": "File",
"mnemonic": "F",
"id": "file",
"children":
[
{ "command": "new_file", "caption": "New File", "mnemonic": "N" },
{ "command": "prompt_open_file", "caption": "Open File…", "mnemonic": "O", "platform": "!OSX" },
{ "command": "prompt_open_folder", "caption": "Open Folder…", "platform": "!OSX" },
{ "command": "prompt_open", "caption": "Open…", "platform": "OSX" },
{
"caption": "Open Recent",
"mnemonic": "R",
"children":
[
{ "command": "reopen_last_file", "caption": "Reopen Closed File" },
{ "caption": "-" },
{ "command": "open_recent_file", "args": {"index": 0 } },
{ "command": "open_recent_file", "args": {"index": 1 } },
{ "command": "open_recent_file", "args": {"index": 2 } },
{ "command": "open_recent_file", "args": {"index": 3 } },
{ "command": "open_recent_file", "args": {"index": 4 } },
{ "command": "open_recent_file", "args": {"index": 5 } },
{ "command": "open_recent_file", "args": {"index": 6 } },
{ "command": "open_recent_file", "args": {"index": 7 } },
{ "caption": "-" },
{ "caption": "More Recent Files",
"mnemonic": "F",
"children":
[
{ "command": "open_recent_file", "args": {"index": 8 } },
{ "command": "open_recent_file", "args": {"index": 9 } },
{ "command": "open_recent_file", "args": {"index": 10 } },
{ "command": "open_recent_file", "args": {"index": 11 } },
{ "command": "open_recent_file", "args": {"index": 12 } },
{ "command": "open_recent_file", "args": {"index": 13 } },
{ "command": "open_recent_file", "args": {"index": 14 } },
{ "command": "open_recent_file", "args": {"index": 15 } },
{ "command": "open_recent_file", "args": {"index": 16 } },
{ "command": "open_recent_file", "args": {"index": 17 } },
{ "command": "open_recent_file", "args": {"index": 18 } },
{ "command": "open_recent_file", "args": {"index": 19 } },
{ "command": "open_recent_file", "args": {"index": 20 } },
]
},
{ "caption": "-" },
{ "command": "open_recent_folder", "args": {"index": 0 } },
{ "command": "open_recent_folder", "args": {"index": 1 } },
{ "command": "open_recent_folder", "args": {"index": 2 } },
{ "command": "open_recent_folder", "args": {"index": 3 } },
{ "command": "open_recent_folder", "args": {"index": 4 } },
{ "command": "open_recent_folder", "args": {"index": 5 } },
{ "command": "open_recent_folder", "args": {"index": 6 } },
{ "command": "open_recent_folder", "args": {"index": 7 } },
{ "caption": "-" },
{ "caption": "More Recent Directories",
"mnemonic": "D",
"children":
[
{ "command": "open_recent_folder", "args": {"index": 8 } },
{ "command": "open_recent_folder", "args": {"index": 9 } },
{ "command": "open_recent_folder", "args": {"index": 10 } },
{ "command": "open_recent_folder", "args": {"index": 11 } },
{ "command": "open_recent_folder", "args": {"index": 12 } },
{ "command": "open_recent_folder", "args": {"index": 13 } },
{ "command": "open_recent_folder", "args": {"index": 14 } },
{ "command": "open_recent_folder", "args": {"index": 15 } },
{ "command": "open_recent_folder", "args": {"index": 16 } },
{ "command": "open_recent_folder", "args": {"index": 17 } },
{ "command": "open_recent_folder", "args": {"index": 18 } },
{ "command": "open_recent_folder", "args": {"index": 19 } },
{ "command": "open_recent_folder", "args": {"index": 20 } },
]
},
{ "caption": "-" },
{ "command": "clear_recent_files", "caption": "Clear Items" }
]
},
... (continued)
在线重建索引
OPTIMIZE TABLE对常规和分区的InnoDB表使用在线DDL，从而减少了并发DML操作的停机时间。由OPTIMIZE TABLE触发并由ALTER TABLE ... FORCE在后台执行的表重建已就位。排他表锁定仅在操作的准备阶段和提交阶段短暂进行。在准备阶段，将更新元数据并创建一个中间表。在提交阶段，将提交表元数据更改。
来源：https://dev.mysql.com/doc/refman/8.0/en/optimize-table.html
但是！我听说过很多事情，您不应该在InnoDB中重建索引，因为它们始终是最新的。 Google对此有所了解。
答案 2 :(得分：0)
pt-online-schema-change可用于优化表。 OPTIMIZE TABLE实际上是noop ALTER TABLE。
pt-online-schema-change --alter "ENGINE=InnoDB" D=sakila,t=actor