When it comes to performance optimization on XFS filesystems, allocation groups are only the beginning. To maintain a high-availability PostgreSQL server, we want to get the most out of XFS. For us, this means using specific mount options.
Thankfully, unlike formatting, mount options can be changed frequently and require very little downtime. Though it isn't essential that we apply these values immediately, the options discussed in this recipe are our recommendation for this stack.
In order to mount an XFS filesystem, we need one to exist. Please follow the recipe contained in Formatting an XFS Filesystem before continuing.
Assuming pg1
is our current primary node, follow these steps as the root
user:
Use this command to find the Linux kernel version:
uname -r
For kernel versions
3.0
and above, mount the filesystem with this command:mount -t xfs -o noatime,nodiratime \-o logbsize=256k,allocsize=1m \/dev/VG_POSTGRES/LV_DATA /db
For kernels...