PostgreSQL Performance Tuning
The physical nature of disk drives makes their performance characteristics different from the other storage areas mentioned in this article. The other storage areas can access any byte with equal speed. Disk drives, with their spinning platters and moving heads, access data near the head's current position much faster than data farther away.
Moving the disk head to another cylinder on the platter takes quite a bit of time. UNIX kernel developers know this. When storing a large file on disk, they try to place the pieces of the file near each other. For example, suppose a file requires ten blocks on disk. The operating system may place blocks 1-5 on one cylinder and blocks 6-10 on another cylinder. If the file is read from beginning to end, only two head movements are required—one to get to the cylinder holding blocks 1-5, and another to get to blocks 6-10. However, if the file is read non-sequentially, e.g., blocks 1,6,2,7,3,8,4,9,5,10; ten head movements are required. As you can see, with disks, sequential access is much faster than random access. This is why PostgreSQL prefers sequential scans to index scans if a significant portion of the table needs to be read. This also highlights the value of the cache.
The disk head moves around quite a bit during database activity. If too many read/write requests are made, the drive can become saturated, causing poor performance (Vmstat and sar can provide information on the amount of activity on each disk drive).
One solution to disk saturation is to move some of the PostgreSQL data files to other disks. Remember, moving the files to other filesystems on the same disk drive does not help. All filesystems on a drive use the same disk heads. Database access can be spread across disk drives in several ways:
Moving Databases—initlocation allows you to create databases on different drives.
Moving Tables—symbolic links allow you to move tables and indexes to different drives. Movement should only be done while PostgreSQL is shut down. Also, PostgreSQL doesn't know about the symbolic links, so if you delete the table and recreate it, it will be created in the default location for that database. In 7.1, pg_database.oid and pg_class.relfilenode map database, table and index names to their numeric filenames.
Moving Indexes—symbolic links allow moving indexes to different drives from their heap tables. This allows an index scan to be performed on one disk while a second disk performs heap lookups.
Moving Joins—symbolic links allow the movement of joined tables to separate disks. If tables A and B are joined, lookups of table A can be performed on one drive while lookups of table B can be done on a second drive.
Moving Log—symbolic links can be used to move the pg_xlog directory to a different disk drive. (Pg_xlog exists in PostgreSQL releases 7.1 and later.) Unlike other writes, PostgreSQL log writes must be flushed to disk before completing a transaction. The cache cannot be used to delay these writes. Having a separate disk for log writes allows the disk head to stay on the current log cylinder so writes can be performed without head movement delay. You can use the postgres -F parameter to prevent log writes from being flushed to disk, but an operating system crash may require a restore from backup.
Other options include the use of RAID features to spread a single filesystem across several drives.
Fortunately, PostgreSQL doesn't require much tuning. Most parameters are automatically adjusted to maintain optimum performance. Cache size and sort size are two parameters administrators can control to make better use of available memory. Disk access can also be spread across drives. Other parameters may be set in share/postgresql.conf.sample. You can copy this file to data/postgresql.conf to experiment with some of PostgreSQL's even more exotic parameters.
|PHP for Non-Developers||Jun 30, 2015|
|A Code Boot Camp for Underprivileged Kids||Jun 30, 2015|
|Comprehensive Identity Management and Audit for Red Hat Enterprise Linux||Jun 29, 2015|
|Linux Kernel 4.1 Released||Jun 26, 2015|
|Secure Server Deployments in Hostile Territory||Jun 25, 2015|
|Take Control of Growing Redis NoSQL Server Clusters||Jun 24, 2015|
- Comprehensive Identity Management and Audit for Red Hat Enterprise Linux
- A Code Boot Camp for Underprivileged Kids
- Linux Kernel 4.1 Released
- Secure Server Deployments in Hostile Territory
- PHP for Non-Developers
- Django Templates
- Cinnamon 2.6 Released
- Gettin' Sticky with It
- Attack of the Drones
- Take Control of Growing Redis NoSQL Server Clusters