[Date Prev][Date Next]
LMDB: fast writes: removing the write mutex ?
- To: email@example.com
- Subject: LMDB: fast writes: removing the write mutex ?
- From: Simon Majou <firstname.lastname@example.org>
- Date: Sun, 13 Apr 2014 18:04:23 +0200
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=majou.org; s=google; h=mime-version:date:message-id:subject:from:to:content-type; bh=DNET0sHoTT66M5l7SXugW0ZE1l/3HasSw9yYcv5aANo=; b=MxhyJ2AN4SbqB+sB1TNjofPcp6W9V4SsehuW6JlhYLA3iadFQr7OOP36H2n3K9hh9Z R5gV1CVZR5G9oY2t2v/ePQTiVOHyQpGpPfR/7wWe5y6ejaXxYg7ve+lJqMo092TmnPky yPyrYoIRBLPGgydmf8ObQDBPUAQeNpQ5CAROc=
I wanted to know, how we could disable the write lock, in case the application shards the writes per thread ? I guess that would imply to have one btree per thread, isn't it ?
We could as well use as many DB as threads (at the cost of also sharding the reads) ?
In that case, how can I disable the write lock ?
Would commenting LOCK_MUTEX_W() & UNLOCK_MUTEX_W() make the trick ?