[DOC] AHI partitions not always optimal

[Kenny Gryp]

See Chat:

10:44 <@gryp> laurynas: hi! about multiple AH partitions. that's per index. How is the size of these partitions? Is
              that in case of 2: 50% or can they grow to different sizes?
10:45 < laurynas> gryp: actually I don't know from the top of my head
10:47 < laurynas> gryp: let's see if I can see this in the source code quickly
10:48 < laurynas> gryp: 2: 50%
10:49 <@gryp> laurynas: auch! OK. Thanks.
10:50 < laurynas> :) yep not very good for very large few tables

[Nilnandan Joshi (nilnandan-joshi)]

Set to confirmed. We should update details in doc that
"hash indexes are assigned to each one based on index_id" is not enough
" When greater than one, the specified number of partitions are created across which to perform the adaptive search."
 in this case we should explain that index is split into N equal parts

http://www.percona.com/doc/percona-server/5.5/scalability/innodb_adaptive_hash_index_partitions.html

[Hrvoje Matijakovic (hrvojem)]

Fixed here: https://github.com/percona/percona-server/pull/382

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PS-1602