Manual:$wgParserCacheType
パーサーのキャッシュ: $wgParserCacheType | |
---|---|
パーサー キャッシュの設定 |
|
導入されたバージョン: | 1.5.0 |
除去されたバージョン: | 使用中 |
許容される値: | (キャッシュ定数のいずれか) 下記参照 |
既定値: | CACHE_ANYTHING |
その他の設定: アルファベット順 | 機能順 |
詳細
The cache type for storing page content HTML (e.g. parsed from wikitext).
Parsing wikitext is considered an expensive operation. It is recommended to give your parser cache plenty of storage space, such that long tail cache hits are possible.
The default parser cache backend (when MainCacheType is left to CACHE_NONE) is effectively CACHE_DB (SqlBagOStuff). If you set up a main cache type such as memcached, it is recommended to set this explicitly to CACHE_DB.
Here are the available cache constants, as defined in Defines.php
CACHE_ANYTHING
- Use anything, as long as it worksCACHE_NONE
- Do not cacheCACHE_DB
- Store cache objects in the DBCACHE_MEMCACHED
- memcached, must specify servers in Manual:$wgMemCachedServersCACHE_ACCEL
- APC, or WinCache, whichever is available, in this order. Determined by the function "ObjectCache::newAccelerator".
注記
大規模ウィキ ファーム向けの助言:
- Consider allocating a dedicated database to ParserCache. Register it in $wgObjectCaches and point $wgParserCacheType to it.
- Consider using MultiWriteBagOStuff to add a higher tier with Memcached in front of the lower database tier.
- Consider setting
'purgePeriod' => 0
in the dedicated SqlBagOStuff entry in $wgObjectCaches. This disables the automatic purging of expired rows (which would normally happen in the background of write requests). You can then schedule the purgeParserCache.php script to e.g. once a day prune expired rows from the a dedicated maintenance server.