Red Hat ENTERPRISE LINUX 3 Reference Manual

Also see for ENTERPRISE LINUX 3 - SECURITY GUIDE: Tuning guide Installation guide Installation guide Installation guide Manual

Contents

Chapter 5. The proc File System 75improve performance on a system using a lot of swap space by telling the kernel to write pages inlarge chunks, minimizing the number of disk seeks.• max_map_count — Configures the maximum number of memory map areas a process may have.In most cases, the default value of 65536 is appropriate.• hugetlb_pool — Specifies in megabytes the size of hugepages—large, physically continuousmemory pages. These pages can be used to create large TLB virtual memory mappings, which areuseful for RAM-intensive database applications. TLB is a special in-CPU cache designed for suchvirtual memory mappings.Although this feature can be adjusted at runtime, it is sensitive to fragmentation issues when grow-ing the pool while the MemFree field of /proc/meminfo is low. For this reason, it may be nec-essary to place new settings in /etc/sysctl.conf and reboot. Refer to Section 5.4 Using thesysctl Command for information about sysctl.Refer to /usr/src/linux-2.4/Documentation/vm/hugetlbpage.txt for more informa-tion about configuring hugepages.ImportantDespite being mentioned in hugetlbpage.txt, the file /proc/sys/vm/nr_hugepages does not ex-ist within Red Hat Enterprise Linux. Instead hugetlb_pool should be used to configure the numberof hugepages available for the system. Also, it is important to remember that nr_hugepages useshugepage units, while hugetlb_pool configures hugepages in megabytes.• overcommit_memory — Configures the conditions under which a large memory request is ac-cepted or denied. The following three modes are available:• 0 — The kernel performs heuristic memory overcommit handling, by estimating the amount ofmemory available and failing requests that are blatantly invalid. Unfortunately, since memoryis allocated using a heuristic rather than a precise algorithm, this setting can sometimes allowoverloading the memory available on a system. This is the default setting.• 1 — The kernel performs no memory overcommit handling. Under this setting, the potential formemory overload is increased, but so is performance for memory intensive tasks (such as thoseexecuted by some scientific software).• 2 — The kernel fails requests for memory that add up to all of swap plus the percent of physi-cal RAM specified in /proc/sys/vm/overcommit_ratio. This setting is best for those whodesire less risk of memory overcommitment.NoteThis setting is only recommended for systems with swap areas larger than physical memory.• overcommit_ratio — Specifies the percentage of physical RAM considered when/proc/sys/vm/overcommit_memory is set to 2. The default value is 50.• pagecache — Controls the percentage of total system memory used for page cache. The followingis a typical output for this file:1 15 100The first value sets the minimum memory to be used as buffer and page cache memory. The middlevalue sets the percentage of system memory dedicated to buffer and page cache memory where thememory management subsystem begins to clear buffer cache more than other kinds of memory tocompensate for a general lack of free memory. The final value specifies the maximum memory tobe used as buffer memory, but is not yet implemented.