Introduction to the use and disabling of transparent huge pages in Linux

introduction

As computing needs continue to grow, applications' demand for memory is also increasing. In order to implement the virtual memory management mechanism, the operating system implements paging management on the memory. Since the memory "paging mechanism" was proposed, the default size of the memory page has been set to 4096 bytes (4KB). Although the memory page size is configurable in principle, the default 4KB page is still used in most operating system implementations. 4KB pages were reasonable when the "paging mechanism" was proposed, because the memory size at that time was only tens of megabytes. However, when the physical memory capacity has grown to several GB or even tens of GB, the operating system still uses 4KB as the basic unit of the page. Is it still reasonable?

When running an application with large memory requirements on the Linux operating system, since the default page size used is 4KB, more TLB Misses and page faults will be generated, which will greatly affect the performance of the application. When the operating system uses 2MB or even larger paging units, the number of TLB Misses and page faults will be greatly reduced, significantly improving application performance. This is also the direct reason why the Linux kernel introduced large page support. The benefits are obvious. Suppose an application requires 2MB of memory. If the operating system uses 4KB as the unit of paging, 512 pages are required, and 512 entries are required in the TLB. At the same time, 512 page table entries are also required. The operating system needs to experience at least 512 TLB Misses and 512 page faults to map the entire 2MB application space to physical memory. However, when the operating system uses 2MB as the basic unit of paging, only one TLB Miss and one page fault are needed to establish a virtual-to-real mapping for the 2MB application space, and no TLB Miss and page fault interrupts are required during operation (assuming that no TLB entry replacement and swap occur).

In order to achieve large page support at the lowest cost, the Linux operating system uses 2M byte large page support based on the hugetlbfs special file system. This method of supporting large pages in the form of a special file system allows applications to flexibly choose the virtual memory page size as needed without being forced to use 2MB large pages.

When Redis is started in Linux, it usually reports WARNING you have Transparent Huge Pages (THP) support enabled in your kernel , which may cause Redis delays and memory usage problems.

Regarding transparent large pages, let’s take a look at the official introduction

Transparent Huge Pages (THP) are enabled by default in RHEL 6 for all applications. The kernel attempts to allocate hugepages whenever possible and any Linux process will receive 2MB pages if the mmap region is 2MB naturally aligned. The main kernel address space itself is mapped with hugepages, reducing TLB pressure from kernel code. For general information on Hugepages, see: What are Huge Pages and what are the advantages of using them?

The kernel will always attempt to satisfy a memory allocation using hugepages. If no hugepages are available (due to non availability of physically continuous memory for example) the kernel will fall back to the regular 4KB pages. THP are also swappable (unlike hugetlbfs). This is achieved by breaking the huge page to smaller 4KB pages, which are then swapped out normally.

But to use hugepages effectively, the kernel must find physically continuous areas of memory big enough to satisfy the request, and also properly aligned. For this, a khugepaged kernel thread has been added. This thread will occasionally attempt to substitute smaller pages being used currently with a hugepage allocation, thus maximizing THP usage.

In userland, no modifications to the applications are necessary (hence transparent). But there are ways to optimize its use. For applications that want to use hugepages, use of posix_memalign() can also help ensure that large allocations are aligned to huge page (2MB) boundaries.

Also, THP is only enabled for anonymous memory regions. There are plans to add support for tmpfs and page cache. THP tunables are found in the /sys tree under /sys/kernel/mm/redhat_transparent_hugepage.

Check whether transparent huge pages are enabled

1: Command cat /sys/kernel/mm/redhat_transparent_hugepage/enabled This command is applicable to Red Hat Enterprise Linux system

[root@getlnx06 ~]# more /etc/issue
 
Red Hat Enterprise Linux Server release 6.6 (Santiago)
 
Kernel \r on an \m
 
[root@getlnx06 ~]# cat /sys/kernel/mm/redhat_transparent_hugepage/enabled
 
[always] madvise never

2: Command cat /sys/kernel/mm/transparent_hugepage/enabled This command is applicable to other Linux systems

[root@getlnx06 ~]# cat /sys/kernel/mm/transparent_hugepage/enabled
 
always madvise [never]
 
[root@getlnx06 ~]#

When you use the command to view, if the output result is [always], it means that transparent huge pages are enabled. [never] means transparent huge pages are disabled, [madvise] means

3: If HugePages_Total returns 0, it means transparent huge pages are disabled.

[root@getlnx06 ~]# grep -i HugePages_Total /proc/meminfo 
 
HugePages_Total: 0

4: cat /proc/sys/vm/nr_hugepages returns 0, which also means transparent huge pages are disabled.

[root@getlnx06 ~]# cat /proc/sys/vm/nr_hugepages 

0

Disable and enable transparent huge pages function

Method 1: Set the /etc/grub.conf file to disable it when the system starts.

[root@getlnx06 ~]# vi /etc/grub.conf
# grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this file
# NOTICE: You have a /boot partition. This means that
# all kernel and initrd paths are relative to /boot/, eg.
# root (hd0,0)
# kernel /vmlinuz-version ro root=/dev/mapper/VolGroup--LogVol0-LogVol01
# initrd /initrd-[generic-]version.img
#boot=/dev/sda
default=0
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title Red Hat Enterprise Linux 6 (2.6.32-504.el6.x86_64)
    root (hd0,0)
    kernel /vmlinuz-2.6.32-504.el6.x86_64 ro root=/dev/mapper/VolGroup--LogVol0-LogVol01 rd_NO_LUKS LANG=en_US.UTF-8 rd_NO_MD SYSFONT=latarcyrheb-sun16 crashkernel=auto rd_LVM_LV=VolGroup-LogVol0/LogVol01 rd_LVM_LV=VolGroup-LogVol0/LogVol00 KEYBOARDTYPE=pc KEYTABLE=us rd_NO_DM rhgb quiet
    initrd /initramfs-2.6.32-504.el6.x86_64.img
transparent_hugepage=never

Method 2: Set up the /etc/rc.local file

[root@getlnx06 ~]# vi /etc/rc.local
#!/bin/sh
#
# This script will be executed *after* all the other init scripts.
# You can put your own initialization stuff in here if you don't
# want to do the full Sys V style init stuff.
touch /var/lock/subsys/local
if test -f /sys/kernel/mm/redhat_transparent_hugepage/enabled; then
  echo never > /sys/kernel/mm/redhat_transparent_hugepage/enabled
fi

After using the above configuration, you must restart the operating system for it to take effect. You can also run the following command without restarting the operating system.

[root@getlnx06 ~]# echo never > /sys/kernel/mm/redhat_transparent_hugepage/enabled
[root@getlnx06 ~]# cat /sys/kernel/mm/redhat_transparent_hugepage/enabled
always madvise [never]
[root@getlnx06 ~]#

Tips:

1: Starting from RedHat 6, OEL 6, SLES 11 and UEK2 kernels, Transparent HugePages is enabled by default: used to improve the performance of memory management. Transparent HugePages is similar to the huge pages in previous versions. The main difference is that Transparent HugePages can be configured in real time and does not require a restart for the configuration to take effect;

2: Transparent Huge Pages are not supported in 32-bit RHEL 6.

3: ORACLE officially does not recommend that you enable Transparent HugePages when using RedHat 6, OEL 6, SLES 11 and UEK2 kernels, because Transparent HugePages has some problems:

• Transparent HugePages in a RAC environment can cause abnormal node restarts and performance issues
• In a stand-alone environment, Transparent HugePages can also cause some unusual performance issues.

Summarize

The above is the full content of this article. I hope that the content of this article will have certain reference learning value for your study or work. Thank you for your support of 123WORDPRESS.COM. If you want to learn more about this, please check out the following links