linux下释放cache内存

本文详细介绍了Linux系统如何管理和释放缓存内存,包括通过调整/proc/sys/vm/drop_caches参数来手动清理缓存的方法。

linux下释放cache内存

细心的朋友会注意到,当你在linux下频繁存取文件后,物理内存会很快被用光,当程序结束后,内存不会被正常释放,而是一直作为caching.这个问题,貌似有不少人在问,不过都没有看到有什么很好解决的办法.那么我来谈谈这个问题.

先来说说free命令

[root@server ~]# free -m
             total       used       free     shared    buffers     cached
Mem:           249        163         86          0         10         94
-/+ buffers/cache:         58        191
Swap:          511          0        511

其中:

total 内存总数

used 已经使用的内存数

free 空闲的内存数

shared 多个进程共享的内存总额

buffers Buffer Cache和cached Page Cache 磁盘缓存的大小

-buffers/cache 的内存数:used - buffers - cached

+buffers/cache 的内存数:free + buffers + cached

可用的memory=free memory+buffers+cached

有了这个基础后,可以得知,我现在used为163MB,free为86,buffer和cached分别为10,94

那么我们来看看,如果我执行复制文件,内存会发生什么变化.

[root@server ~]# cp -r /etc ~/test/
[root@server ~]# free -m
             total       used       free     shared    buffers     cached
Mem:           249        244          4          0          8        174
-/+ buffers/cache:         62        187
Swap:          511          0        511

在我命令执行结束后,used为244MB,free为4MB,buffers为8MB,cached为174MB,天呐都被cached吃掉了.别紧张,这是为了提高文件读取效率的做法.

引用[url]http://www.2qyou.com/thread-591-1-1.html[/url] 为了提高磁盘存取效率, Linux做了一些精心的设计, 除了对dentry进行缓存(用于VFS,加速文件路径名到inode的转换), 还采取了两种主要Cache方式:Buffer Cache和Page Cache。前者针对磁盘块的读写,后者针对文件inode的读写。这些Cache有效缩短了 I/O系统调用(比如read,write,getdents)的时间。"

那么有人说过段时间,linux会自动释放掉所用的内存,我们使用free再来试试,看看是否有释放>?

[root@server test]# free -m
             total       used       free     shared    buffers     cached
Mem:           249        244          5          0          8        174
-/+ buffers/cache:         61        188
Swap:          511          0        511

MS没有任何变化,那么我能否手动释放掉这些内存呢???回答是可以的!

/proc是一个虚拟文件系统,我们可以通过对它的读写操作做为与kernel实体间进行通信的一种手段.也就是说可以通过修改/proc中的文件,来对当前kernel的行为做出调整.那么我们可以通过调整/proc/sys/vm/drop_caches来释放内存.操作如下:

[root@server test]# cat /proc/sys/vm/drop_caches
0
首先,/proc/sys/vm/drop_caches的值,默认为0

[root@server test]# sync

手动执行sync命令(描述:sync 命令运行 sync 子例程。如果必须停止系统,则运行 sync 命令以确保文件系统的完整性。sync 命令将所有未写的系统缓冲区写到磁盘中,包含已修改的 i-node、已延迟的块 I/O 和读写映射文件)

[root@server test]# echo 3 > /proc/sys/vm/drop_caches
[root@server test]# cat /proc/sys/vm/drop_caches
3

将/proc/sys/vm/drop_caches值设为3

[root@server test]# free -m
             total       used       free     shared    buffers     cached
Mem:           249         66        182          0          0         11
-/+ buffers/cache:         55        194
Swap:          511          0        511

再来运行free命令,发现现在的used为66MB,free为182MB,buffers为0MB,cached为11MB.那么有效的释放了buffer和cache.

有关/proc/sys/vm/drop_caches的用法在下面进行了说明

/proc/sys/vm/drop_caches (since Linux 2.6.16)
Writing  to  this  file  causes the kernel to drop clean caches,
dentries and inodes from memory, causing that memory  to  become
free.

To  free  pagecache,  use  echo 1 > /proc/sys/vm/drop_caches; to
free dentries and inodes, use echo 2 > /proc/sys/vm/drop_caches;
to   free   pagecache,   dentries  and  inodes,  use  echo  3  >
/proc/sys/vm/drop_caches.

Because this is a non-destructive operation  and  dirty  objects
are not freeable, the user should run sync(8) first.
<think> Hmm, the user is asking about methods to release cache in Linux systems. Looking at the references they provided, it's clear they've already found some technical commands but want a comprehensive understanding. The references mention using echo commands to write to /proc/sys/vm/drop_caches with values 1,2,3. References [1] and [2] both show this approach but [1] gives more context about what each value does while [2] is more concise. I notice the user specifically says "release cache" but technically in Linux we're talking about freeing up memory used by disk buffers (buff) and page cache (cache). The distinction matters because: - Buffers (buff) hold metadata about files/directories - Page cache (cache) stores actual file contents The safest method is starting with echo 1 which only clears page cache. Echo 3 is more aggressive and should be used with caution. The user might be experiencing high memory usage showing as buff/cache in free command output. They probably want to free this memory without rebooting. Important caveats to mention: - This doesn't harm running applications - Kernel will repopulate caches as needed - root privileges required For completeness, I should also mention the sync command option before clearing caches to flush unwritten data to disk. While not strictly necessary, it's good practice. The user seems technical since they're using root commands, so I can include the detailed explanations about what each echo value does. They'll appreciate understanding the underlying mechanism rather than just commands. </think> 在Linux系统中释放缓冲(buff)和缓存(cache)内存,主要可通过以下方法实现: ### 一、手动释放缓存(推荐方法) 执行三条核心命令,需**root权限**: ```bash # 1. 先同步文件系统(确保数据安全) sync # 2. 清除页缓存(PageCache) echo 1 > /proc/sys/vm/drop_caches # 3. 清除目录项和inode缓存 echo 2 > /proc/sys/vm/drop_caches # 4. 同时清除页缓存和inode缓存 echo 3 > /proc/sys/vm/drop_caches ``` **效果验证**: ```bash free -h # 查看内存变化 total used free shared buff/cache available Mem: 7.6G 627M 6.9G 15M 121M 6.8G ``` ### 二、内核参数自动控制 修改`/etc/sysctl.conf`实现自动回收: ```bash # 添加以下参数 vm.vfs_cache_pressure=100 # 提高inode缓存回收优先级(默认100) vm.swappiness=10 # 降低交换分区使用倾向(建议值) ``` 生效配置: ```bash sysctl -p ``` ### 注意事项 1. **操作影响**: - `echo 1`:仅清理文件页缓存,**最安全** - `echo 3`:清除所有缓存,可能导致短暂I/O上升 - 生产环境建议夜间执行 2. **内存机制**: - Buffers:内核块设备缓冲区 - Cache:文件系统页缓存 - Linux自动回收空闲缓存,**手动清理仅用于特殊场景**[^1]
评论 6
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值