Comments on: read_ahead (disabled) as steroid

By: Vadim

Vadim — Thu, 14 May 2009 03:00:10 +0000

Andy,

I do not think so, it depends on IO subsystem. For single HDD read_ahead may be useful.

By: Andy

Andy — Wed, 13 May 2009 22:36:26 +0000

Should read_ahead always be disabled then to improve performance?

What are the pros and cons?

By: Vadim

Vadim — Wed, 13 May 2009 06:27:18 +0000

Mark,

Here is comment from Yasufumi:

- “pause” (rep; nop) instruction in spin loop should increase performance of
mutex/rw_lock contention. It may affect 5~10% in throughput.

- log_buffer_flush_maybe_sync() will decrease the number of log sync
at innodb_flush_log_at_trx_commit=[0|2]. It may affect to performance much.

- The number of request slots for the each IO thread has been increased. (x4)
It might be short for some high-end environments.

- relax condition to flush log and contract ibuf at inner loop of srv_master_thread()
– if (n_pend_ios < 3 && (n_ios – n_ios_old < PCT_IO(5))) {
+ if (n_pend_ios < PCT_IO(3) && (n_ios – n_ios_old < PCT_IO(5))) {

By: Mark Callaghan

Mark Callaghan — Wed, 13 May 2009 06:01:01 +0000

What did you backport from 5.4?