Hello, I'm hitting a performance wall on my MySQL primarily I believe because the bin log and the InnoDB logs are on the same volume group as another MySQL server. In reality, I have four MySQL servers, two per server (in Solaris Containers). All four are sharing the same volume group to maximize the spindle count. Unfortunately, that's driving the seek time crazy.
Optimally, I would give each MySQL server two volume groups (one for its logs and another for its databases). That would ensure that the logs and databases each had dedicated disks in the array. Unfortunately, given my array configuration I'm limited to 6 volume groups (unlimited LUNs inside those VGs), and would need 8 VGs to give each of the 4 servers two VGs. The issue is complicated by the fact that the second 2 MySQL servers are slaves of the first two. So the data written to one server is identically written a few seconds later to the slave. So my question is, which would be the better trade-off: 1.) Put the logs for two master MySQL servers on one VG, and then the databases for those master MySQL servers on a second VG. or 2.) Put the logs for a master and slave MySQL server on the same VG, and then put the databases for the slave and master on a second VG. Or is there a better way of splitting the IO for different disks? Any help is greatly appreciated. Best Regards, Jason -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
