Обсуждение: Raid 10 settings for optimal postgres performance?

I have just configured our disks to raid10 from raid5.
The raid 10 is now 6 drives giving us 750G use by OS and postgres.
 
What is the preferred setup for postgres concerning cache settings on the controller and disks and the preferred block size allocated for the disks when postgres uses block size 8192?
 
df -h
/dev/sda3             750G   65G  647G  10% /
/dev/sda1             496M   17M  454M   4% /boot

show all;
server_version        | 8.4.2
block_size            | 8192
 
 
Controller settings
Read-cache mode - Enabled
Write-cache mode - Enabled (write-back)
Write-cache setting - Enabled (write-back) when protected by battery
Stripe-unit size                         : 256 KB
 
Disk setting
Write Cache                        : Disabled (write-through)
 
Postgresql.conf
fsync = on                              # turns forced synchronization on or off
#synchronous_commit = on                # immediate fsync at commit
wal_sync_method = fsync         # the default is the first option
 
Logical drive number 1
   Logical drive name                       : RAID10
   RAID level                               : 10
   Status of logical drive                  : Okay
   Size                                     : 858000 MB
   Read-cache mode                          : Enabled
   Write-cache mode                         : Enabled (write-back)
   Write-cache setting                      : Enabled (write-back) when protected by battery
   Partitioned                              : Yes
   Number of segments                       : 6
   Stripe-unit size                         : 256 KB
   Stripe order (Channel,Device)            : 0,0 0,1 0,2 0,3 0,4 0,5
   Defunct segments                         : No
   Defunct stripes                          : No

 
      Device #0
         Device is a Hard drive
         State                              : Online
         Supported                          : Yes
         Transfer Speed                     : SAS 3.0 Gb/s
         Reported Channel,Device            : 0,0
         Vendor                             : IBM-ESXS
         Model                              : ST3300655SS
         Firmware                           : BA26
         World-wide name                    : 5000C5000A42EFAC
         Size                               : 286102 MB
         Write Cache                        : Disabled (write-through)
         FRU                                : None
         PFA
 
Sincerely,
 
Chris
 

---

Stay in touch. Get Messenger on your phone now.

Chris Barnes wrote:

What is the preferred setup for postgres concerning cache settings on the controller and disks and the preferred block size allocated for the disks when postgres uses block size 8192? 
 
Controller settings
Read-cache mode - Enabled
Write-cache mode - Enabled (write-back)
Write-cache setting - Enabled (write-back) when protected by battery
Stripe-unit size                         : 256 KB
 
Disk setting
Write Cache                        : Disabled (write-through) 

You already have the right cache settings:  write-back when protected by battery, write cache on disks disabled.  And there aren't any interesting settings to change for block of stripe size on a RAID10 array that do anything you'll be able to measure.  You're done with controller and disk setup.

Looks like a standard ext3 filesystem under Linux, right?  Next you should run tests with bonnie++ to make sure the array is performing as expecting.

-- 
Greg Smith  2ndQuadrant US  Baltimore, MD
PostgreSQL Training, Services and Support
greg@2ndQuadrant.com   www.2ndQuadrant.us