For the innodb_log_block_size variable, the manual states:

"Size in bytes of the transaction log records. Generally 512, the default, or 4096, for SSD cards, are the only two useful values. If the server is restarted and this value is changed, all old log files need to be removed. Should be set to 4096 if innodb_flush_method is set to ALL_O_DIRECT on ext4 filesystems."

https://mariadb.com/kb/en/mariadb/documentation/storage-engines/xtradb-and-innodb/xtradbinnodb-server-system-variables/#innodb_log_block_size

What I find unclear is the last statement, "Should be set to 4096 if innodb_flush_method is set to ALL_O_DIRECT on ext4 filesystems."

Does this mean if you still are using SSD cards or not?

To me, it sounds as if this is separate from the SSD cards.

If so, I think that should be made more clear, as that could be a good optimization for those who run with ALL_O_DIRECT on ext4.

Otherwise, if it only applies to SSD, I think the working should reflect that.

For what it's worth, I read about the variable in the Percona manual too, and it states:

"This variable changes the size of transaction log records. The default size of 512 bytes is good in most situations. However, setting it to 4096 may be a good optimization with SSD cards. While settings other than 512 and 4096 are possible, as a practical matter these are really the only two that it makes sense to use. Clean restart and removal of the old logs is needed for the variable innodb_log_block_size to be changed."

http://www.percona.com/doc/percona-server/5.5/scalability/innodb_io_55.html#innodb_log_block_size

So again, it does not sound like the "ALL_O_DIRECT on ext4" applies to SSD card usage. However, there is no mention of "ALL_O_DIRECT on ext4", so I think we need further clarification.