Where did the vm.pagecache parameter go and how to replace it?

Question

linux

What went into / dev / null - I already know There was a vm.pagecache parameter that limits the amount of memory that is allowed to be occupied by the cache. The default is 100 (%), reducing it is simply necessary when virtualizing (either on a VPS or in your hypervisor - everywhere) and when working without a swap.
In newer kernels (2.6.35, 3.x), when setting vm.pagecache in sysctl.conf, an error is thrown, and ls /proc/sys/vm/ says that there is no such parameter. A question for connoisseurs of new Linux distributions: how, how to restore lost control over memory allocation?
Yes, and drop_caches in the crown is not the way.

Answer 1

Look towards vfs_cache_pressure .
More recently, I asked a very similar question . As a result, I came to the swappines and vfs_cache_pressure parameters. In my case (a large volume of tmpfs), this did not help much and I had to turn on swap. But the meaning is similar to what you are looking for, although not exactly. If you find the exact analogue of pagecache - please write. It would be very helpful.

Answer 2

Set vfs_cache_pressure greater than 100
_ At the default value of vfs_cache_pressure=100 the kernel will attempt to reclaim dentries and inodes at a "fair" rate with respect to pagecache and swapcache reclaim. Decreasing vfs_cache_pressure causes the kernel to prefer to retain dentry and inode caches. When vfs_cache_pressure=0, the kernel will never reclaim dentries and inodes due to memory pressure and this can easily lead to out-of-memory conditions. Increasing vfs_cache_pressure beyond 100
causes the kernel to prefer to reclaim dentries and inodes.