OS: Tuning Manajemen Memory - Revision history

Onnowpurbo: /* Monitoring Perilaku VM */

2014-04-05T21:17:12Z

Monitoring Perilaku VM

← Older revision		Revision as of 21:17, 5 April 2014
Line 147:		Line 147:
	* /proc/meminfo: File ini memberikan detail kemana saja memory digunakan.		* /proc/meminfo: File ini memberikan detail kemana saja memory digunakan.

	* slabtop: tool ini memberikan informasi detail tentang penggunaan memory slab ~~kerne~~. buffer_head, dentry, inode_cache, ext3_inode_cache, dll adalah cache utama. Perintah ini merupakan bagian dari paket procps.		* slabtop: tool ini memberikan informasi detail tentang penggunaan memory slab kernel. buffer_head, dentry, inode_cache, ext3_inode_cache, dll adalah cache utama. Perintah ini merupakan bagian dari paket procps.

	==Referensi==		==Referensi==

Onnowpurbo: /* Menguragi Penggunaan Memory */

2013-06-18T03:12:42Z

Menguragi Penggunaan Memory

← Older revision		Revision as of 03:12, 18 June 2013
Line 53:		Line 53:
	Ini adalah in-memory cache dari directory entry di sistem. Ini berisi nama (nama file), dan inode yang nunjuk pada file tersebut, dan children entry. Cache ini digunakan saat kita mengakses struktur directiry dan mengakses file menggunakan nama.		Ini adalah in-memory cache dari directory entry di sistem. Ini berisi nama (nama file), dan inode yang nunjuk pada file tersebut, dan children entry. Cache ini digunakan saat kita mengakses struktur directiry dan mengakses file menggunakan nama.

	==~~Menguragi~~ Penggunaan Memory==		==Mengurangi Penggunaan Memory==

	===Mengurangi Penggunaan malloc (Anonymous)===		===Mengurangi Penggunaan malloc (Anonymous)===

Onnowpurbo: /* Referensi */

2013-05-08T22:55:34Z

Referensi

← Older revision		Revision as of 22:55, 8 May 2013
Line 152:		Line 152:

	* http://doc.opensuse.org/documentation/html/openSUSE/opensuse-tuning/cha.tuning.memory.html		* http://doc.opensuse.org/documentation/html/openSUSE/opensuse-tuning/cha.tuning.memory.html
			* http://www.pythian.com/blog/the-mysterious-world-of-shmmax-and-shmall/

	==Pranala Menarik==		==Pranala Menarik==

Onnowpurbo: /* Pranala Menarik */

2013-04-07T06:44:21Z

Pranala Menarik

@@ Line 157: / Line 157: @@
 * [[Linux]]
 * [[Ubuntu]]
 * [[Sistem Operasi]]
 * [[Linux: Instalasi Sistem Operasi]]
-* [[Linux: Skema Partisi di Linux]]
 * [[Kernel]]
 * [[Compile Kernel]]
 * [[Compile Kernel: Konfigurasi Kernel]]
-* [[Kernel: Anatomi Kernel Source]]
-* [[OS: Linux Kernel]]
+===Remaster Linux===
 * [[OS: Parameter Kernel Default]]
 * [[OS: Kernel Scheduler]]
 * [[OS: Complete Teori Tuning Kernel Scheduler]]
 * [[OS: Complete Teori Tuning I/O Performance]]
 * [[OS: Tuning Manajemen Memory]]
-* [[OS: Tuning Kernel Scheduler]]
-* [[OS: Tuning Completely Fair Queueing CFQ I/O scheduler]]
+===Android===
-* [[OS: Tuning Completely Fair scheduler CFS]]
 * [[OS: Build in Monitoring Tool]]
 * [[Linux Benchmarking]]
 * [[OS: Benchmarking menggunakan UnixBench]]
 * [[OS: Benchmarking menggunakan LLCBench]]

Onnowpurbo: /* Non-Uniform Memory Access (NUMA) */

2013-04-03T03:57:17Z

Non-Uniform Memory Access (NUMA)

← Older revision		Revision as of 03:57, 3 April 2013
Line 131:		Line 131:
	==Non-Uniform Memory Access (NUMA)==		==Non-Uniform Memory Access (NUMA)==

	~~Another increasingly important role of the~~ VM ~~is to provide good~~ NUMA ~~allocation strategies~~. NUMA ~~stands for~~ non-uniform memory access, ~~and most of today's~~ multi-socket ~~servers are~~ NUMA ~~machines~~. NUMA ~~is a secondary concern to managing~~ swapping ~~and caches in terms of~~ performance~~, and there are lots of documents about improving~~ NUMA ~~memory allocations~~. ~~One particular~~ parameter ~~interacts with~~ page ~~reclaim~~:		Salah satu fungsi VM yang semakin penting adalah memberikan strategi alokasi NUMA yang baik. NUMA singkatan dari non-uniform memory access, dan sebagian besar multi-socket server hari ini adalah mesin NUMA. NUMA adalah perhatian kedua dalam mengatur swapping dan cache khususnya untuk performance. Aada banyak dokumen yang menjelaskan tentang bagaimana memperbaiki alokasi memory NUMA. Salah satu parameter yang berinteraksi dengan reclaim page adalah:

	'''/proc/sys/vm/zone_reclaim_mode'''		'''/proc/sys/vm/zone_reclaim_mode'''

	~~This parameter controls whether~~ memory reclaim ~~is performed on a local~~ NUMA ~~node even if there is plenty of~~ memory free ~~on other nodes~~. ~~This parameter is automatically turned on on machines with more pronounced~~ NUMA ~~characteristics~~.		Parameter ini mengontrol apakah memory reclaim di lakukan di node lokal NUMA meskipun banyak memory free di node lainnya. Parameter ini secara automatis akan diaktifkan pada mesin yang mempunya NUMA karakteristik yang menonjol.

	~~If the~~ VM ~~caches are not being allowed to fill all of~~ memory ~~on a~~ NUMA ~~machine~~, ~~it could be due to~~ zone_reclaim_mode ~~being~~ set. ~~Setting to~~ 0 ~~will disable this behavior~~.		Jika cache VM tidak di ijinkan untuk mengisi semua memory di mesin NUMA, itu kemungkinan karena zone_reclaim_mode di set. Menset ke 0 akan mematikan perlaku ini.

	==Monitoring Perilaku VM==		==Monitoring Perilaku VM==

Onnowpurbo: /* Parameter Readahead */

2013-04-03T03:51:27Z

Parameter Readahead

← Older revision		Revision as of 03:51, 3 April 2013
Line 119:		Line 119:
	'''/sys/block/<bdev>/queue/read_ahead_kb'''		'''/sys/block/<bdev>/queue/read_ahead_kb'''

	~~If one or more processes are sequentially reading a~~ file, ~~the~~ kernel ~~reads some~~ data ~~in advance~~ (ahead) ~~in order to reduce the amount of time that processes have to wait for~~ data ~~to be available~~. ~~The actual amount of~~ data ~~being read in advance is computed dynamically~~, ~~based on how much~~ "~~sequential~~" ~~the~~ I/O ~~seems to be~~. ~~This parameter sets the maximum amount of~~ data ~~that the~~ kernel ~~reads ahead for a single~~ file. ~~If you observe that large sequential reads from a~~ file ~~are not fast enough~~, ~~you can try increasing this value~~. ~~Increasing it too far may result in~~ readahead ~~thrashing where~~ pagecache ~~used for~~ readahead ~~is reclaimed before it can be used~~, ~~or slowdowns due to a large amount of useless~~ I/O. ~~The~~ default ~~value is~~ 512 (kb).		Jika satu atau lebih proses secara berurutan membaca (read) sebuah file, kernel akan membaca beberapa data lebih dulu (ahead) untuk mengurangi waktu tunggu proses agar data tersedia. Banyaknya data yang dibaca terlebih dulu di hitung secara dinamik, tergantung seberapa "urut" proses I/O tersebut. Parameter ini menset jumlah data maksimum yang dapat di baca oleh kernel dari sebuah file. Jika kita melihat pembacaan berurut dari sebuah file tidak cukup cepat, kita dapat mencoba menaikan nilai ini. Menaikan nilai ini terlalu besar akan menyebabkan readahead membuat sampah pada pagecache yang digunakan, karena readahead akan di reclaim sebelum digunakan, atau menjadi lambat karena jumlah I/O yang percuma. Nilai default adalah 512 (kb).

	===Lebih Lanjut Tentang Parameter VM===		===Lebih Lanjut Tentang Parameter VM===

Onnowpurbo: /* Parameter Writeback */

2013-04-03T03:44:23Z

Parameter Writeback

← Older revision		Revision as of 03:44, 3 April 2013
Line 107:		Line 107:
	'''/proc/sys/vm/dirty_background_ratio'''		'''/proc/sys/vm/dirty_background_ratio'''

	~~This is the percentage of the~~ total ~~amount of~~ free ~~and~~ reclaimable memory. ~~When the amount of~~ dirty ~~pagecache exceeds this percentage~~, writeback ~~threads start writing~~ back dirty ~~memory~~. ~~The~~ default ~~value is~~ 10 (%).		Ini ada persentase dari total free dan reclaimable memory. Jika jumlah dari pagecache dirty melebihi persentase ini, thread writeback akan mulai write back memory dirty. Nilai default adalah 10 (%).

	'''/proc/sys/vm/dirty_ratio'''		'''/proc/sys/vm/dirty_ratio'''

	~~Similar percentage value as above~~. ~~When this is exceeded~~, ~~applications that want to~~ write ~~to the~~ pagecache ~~are blocked and start performing~~ writeback ~~as well~~. ~~The~~ default ~~value is~~ 40 (%).		Sama denga nilai persentase di ats. Jika nilai ini dilalui, aplikasi yang akan write ke pagecache akan di blokir dan mulai melakukan writeback juga. Nilai default adalah 40 (%).

	~~These two values together determine the~~ pagecache writeback ~~behavior~~. ~~If these values are increased~~, ~~more~~ dirty memory ~~is kept in the system for a longer time. With more~~ dirty ~~memory allowed in the system~~, ~~the chance to improve~~ throughput ~~by avoiding writeback~~ I/O ~~and to submitting more optimal~~ I/O ~~patterns increases~~. ~~However~~, ~~more~~ dirty ~~memory can either harm~~ latency ~~when~~ memory ~~needs to be reclaimed or at~~ data ~~integrity~~ (sync) ~~points when it needs to be written back to~~ disk.		Kedua nilai ini secara bersama menentukan perilaku pagecache writeback. Jika nilai ini di naikan, akan lebih banyak memory dirty yang di simpan di sistem untuk waktu yang lebih lama. Dengan lebih banyak memory dirty yang ada di sistem, kemungkinan untuk memperbaiki throughput dengan menghindari I/O writeback dan memberikan pola I/O yang lebih baik akan meningkat. Akan tetapi, dengan lebih banyak memory dirty akan menyebabkan latency ketika memory perlu di reclaim atau saat integritas data (sync) perlu di tulis kembali ke disk.

	===Parameter Readahead===		===Parameter Readahead===

Onnowpurbo: /* Parameter Writeback */

2013-04-03T03:32:02Z

Parameter Writeback

← Older revision		Revision as of 03:32, 3 April 2013
Line 101:		Line 101:
	===Parameter Writeback===		===Parameter Writeback===

	~~One important change in~~ writeback ~~behavior since openSUSE 10 is that modification to file-backed~~ mmap() memory ~~is accounted immediately as~~ dirty ~~memory~~ (~~and subject to~~ writeback). ~~Whereas previously it would only be subject to~~ writeback ~~after it was~~ unmapped, ~~upon an~~ msync() ~~system call~~, ~~or under heavy~~ memory ~~pressure~~.		Perubahan perilaku writeback terutama ke memory mmap() akan di lihat langsung sebagai memory dirty (yang akan diperhitungkan untuk di writeback). Sedang sebelumnya, hanya akan di perhitungkan untuk di writeback, sesudah di unmapped, oleh sistem call msync(), atau di bawah tekanan memory yang sangat berat.

	~~Some applications do not expect~~ mmap ~~modifications to be subject to such~~ writeback ~~behavior~~, ~~and~~ performance ~~can be reduced~~. Berkeley DB (~~and applications using it~~) ~~is one known example that can cause problems~~. ~~Increasing~~ writeback ~~ratios and times can improve this type of~~ slowdown.		Beberapa aplikasi tidak berharap adanya modifikasi mmap dan mengalami perilaku writeback yang baru, dan performance dapat berkurang. Berkeley DB (dan aplikasi yang menggunakannya) adalah salah satu contoh yang diketahui dapat menyebabkan masalah. Menaikan perbandingan dan waktu writeback dan akan memperbaiki slowdown yang seperti ini.

	'''/proc/sys/vm/dirty_background_ratio'''		'''/proc/sys/vm/dirty_background_ratio'''

Onnowpurbo: /* 15.3.4. Further VM Parameters */

2013-04-03T03:24:17Z

15.3.4. Further VM Parameters

← Older revision		Revision as of 03:24, 3 April 2013
Line 121:		Line 121:
	If one or more processes are sequentially reading a file, the kernel reads some data in advance (ahead) in order to reduce the amount of time that processes have to wait for data to be available. The actual amount of data being read in advance is computed dynamically, based on how much "sequential" the I/O seems to be. This parameter sets the maximum amount of data that the kernel reads ahead for a single file. If you observe that large sequential reads from a file are not fast enough, you can try increasing this value. Increasing it too far may result in readahead thrashing where pagecache used for readahead is reclaimed before it can be used, or slowdowns due to a large amount of useless I/O. The default value is 512 (kb).		If one or more processes are sequentially reading a file, the kernel reads some data in advance (ahead) in order to reduce the amount of time that processes have to wait for data to be available. The actual amount of data being read in advance is computed dynamically, based on how much "sequential" the I/O seems to be. This parameter sets the maximum amount of data that the kernel reads ahead for a single file. If you observe that large sequential reads from a file are not fast enough, you can try increasing this value. Increasing it too far may result in readahead thrashing where pagecache used for readahead is reclaimed before it can be used, or slowdowns due to a large amount of useless I/O. The default value is 512 (kb).

	===~~15.3.4. Further~~ VM ~~Parameters~~===		===Lebih Lanjut Tentang Parameter VM===

	~~For the complete list of the~~ VM ~~tunable parameters~~, ~~see~~ /usr/src/linux/Documentation/sysctl/vm.txt ~~(available after having installed the~~ kernel-source ~~package)~~.		Untuk daftar lengkap dari parameter VM yang dapat di tune, ada baiknya melihat isi file

			/usr/src/linux/Documentation/sysctl/vm.txt

			yang tersedia juga anda menginstalasi source code kernel, atau pakey kernel-source.

	==Non-Uniform Memory Access (NUMA)==		==Non-Uniform Memory Access (NUMA)==

Onnowpurbo: /* 15.3.3. Readahead parameters */

2013-04-03T03:22:15Z

15.3.3. Readahead parameters

← Older revision		Revision as of 03:22, 3 April 2013
Line 115:		Line 115:
	These two values together determine the pagecache writeback behavior. If these values are increased, more dirty memory is kept in the system for a longer time. With more dirty memory allowed in the system, the chance to improve throughput by avoiding writeback I/O and to submitting more optimal I/O patterns increases. However, more dirty memory can either harm latency when memory needs to be reclaimed or at data integrity (sync) points when it needs to be written back to disk.		These two values together determine the pagecache writeback behavior. If these values are increased, more dirty memory is kept in the system for a longer time. With more dirty memory allowed in the system, the chance to improve throughput by avoiding writeback I/O and to submitting more optimal I/O patterns increases. However, more dirty memory can either harm latency when memory needs to be reclaimed or at data integrity (sync) points when it needs to be written back to disk.

	===~~15.3.3.~~ Readahead ~~parameters~~===		===Parameter Readahead===

	'''/sys/block/<bdev>/queue/read_ahead_kb'''		'''/sys/block/<bdev>/queue/read_ahead_kb'''

	If one or more processes are sequentially reading a file, the kernel reads some data in advance (ahead) in order to reduce the amount of time that processes have to wait for data to be available. The actual amount of data being read in advance is computed dynamically, based on how much "sequential" the I/O seems to be. This parameter sets the maximum amount of data that the kernel reads ahead for a single file. If you observe that large sequential reads from a file are not fast enough, you can try increasing this value. Increasing it too far may result in readahead thrashing where pagecache used for readahead is reclaimed before it can be used, or slowdowns due to a large amount of useless I/O. The default value is 512 (kb).		If one or more processes are sequentially reading a file, the kernel reads some data in advance (ahead) in order to reduce the amount of time that processes have to wait for data to be available. The actual amount of data being read in advance is computed dynamically, based on how much "sequential" the I/O seems to be. This parameter sets the maximum amount of data that the kernel reads ahead for a single file. If you observe that large sequential reads from a file are not fast enough, you can try increasing this value. Increasing it too far may result in readahead thrashing where pagecache used for readahead is reclaimed before it can be used, or slowdowns due to a large amount of useless I/O. The default value is 512 (kb).

	===15.3.4. Further VM Parameters===		===15.3.4. Further VM Parameters===