Documentation/device-mapper/cache.txt - maze/linux - Git at Google

 Introduction
 ============

 dm-cache is a device mapper target written by Joe Thornber, Heinz
 Mauelshagen, and Mike Snitzer.

 It aims to improve performance of a block device (eg, a spindle) by
 dynamically migrating some of its data to a faster, smaller device
 (eg, an SSD).

 This device-mapper solution allows us to insert this caching at
 different levels of the dm stack, for instance above the data device for
 a thin-provisioning pool.  Caching solutions that are integrated more
 closely with the virtual memory system should give better performance.

 The target reuses the metadata library used in the thin-provisioning
 library.

 The decision as to what data to migrate and when is left to a plug-in
 policy module.  Several of these have been written as we experiment,
 and we hope other people will contribute others for specific io
 scenarios (eg. a vm image server).

 Glossary
 ========

   Migration -  Movement of the primary copy of a logical block from one
 	       device to the other.
   Promotion -  Migration from slow device to fast device.
   Demotion  -  Migration from fast device to slow device.

 The origin device always contains a copy of the logical block, which
 may be out of date or kept in sync with the copy on the cache device
 (depending on policy).

 Design
 ======

 Sub-devices
 -----------

 The target is constructed by passing three devices to it (along with
 other parameters detailed later):

 1. An origin device - the big, slow one.

 2. A cache device - the small, fast one.

 3. A small metadata device - records which blocks are in the cache,
    which are dirty, and extra hints for use by the policy object.
    This information could be put on the cache device, but having it
    separate allows the volume manager to configure it differently,
    e.g. as a mirror for extra robustness.

 Fixed block size
 ----------------

 The origin is divided up into blocks of a fixed size.  This block size
 is configurable when you first create the cache.  Typically we've been
 using block sizes of 256k - 1024k.

 Having a fixed block size simplifies the target a lot.  But it is
 something of a compromise.  For instance, a small part of a block may be
 getting hit a lot, yet the whole block will be promoted to the cache.
 So large block sizes are bad because they waste cache space.  And small
 block sizes are bad because they increase the amount of metadata (both
 in core and on disk).

 Writeback/writethrough
 ----------------------

 The cache has two modes, writeback and writethrough.

 If writeback, the default, is selected then a write to a block that is
 cached will go only to the cache and the block will be marked dirty in
 the metadata.

 If writethrough is selected then a write to a cached block will not
 complete until it has hit both the origin and cache devices.  Clean
 blocks should remain clean.

 A simple cleaner policy is provided, which will clean (write back) all
 dirty blocks in a cache.  Useful for decommissioning a cache.

 Migration throttling
 --------------------

 Migrating data between the origin and cache device uses bandwidth.
 The user can set a throttle to prevent more than a certain amount of
 migration occuring at any one time.  Currently we're not taking any
 account of normal io traffic going to the devices.  More work needs
 doing here to avoid migrating during those peak io moments.

 For the time being, a message "migration_threshold <#sectors>"
 can be used to set the maximum number of sectors being migrated,
 the default being 204800 sectors (or 100MB).

 Updating on-disk metadata
 -------------------------

 On-disk metadata is committed every time a REQ_SYNC or REQ_FUA bio is
 written.  If no such requests are made then commits will occur every
 second.  This means the cache behaves like a physical disk that has a
 write cache (the same is true of the thin-provisioning target).  If
 power is lost you may lose some recent writes.  The metadata should
 always be consistent in spite of any crash.

 The 'dirty' state for a cache block changes far too frequently for us
 to keep updating it on the fly.  So we treat it as a hint.  In normal
 operation it will be written when the dm device is suspended.  If the
 system crashes all cache blocks will be assumed dirty when restarted.

 Per-block policy hints
 ----------------------

 Policy plug-ins can store a chunk of data per cache block.  It's up to
 the policy how big this chunk is, but it should be kept small.  Like the
 dirty flags this data is lost if there's a crash so a safe fallback
 value should always be possible.

 For instance, the 'mq' policy, which is currently the default policy,
 uses this facility to store the hit count of the cache blocks.  If
 there's a crash this information will be lost, which means the cache
 may be less efficient until those hit counts are regenerated.

 Policy hints affect performance, not correctness.

 Policy messaging
 ----------------

 Policies will have different tunables, specific to each one, so we
 need a generic way of getting and setting these.  Device-mapper
 messages are used.  Refer to cache-policies.txt.

 Discard bitset resolution
 -------------------------

 We can avoid copying data during migration if we know the block has
 been discarded.  A prime example of this is when mkfs discards the
 whole block device.  We store a bitset tracking the discard state of
 blocks.  However, we allow this bitset to have a different block size
 from the cache blocks.  This is because we need to track the discard
 state for all of the origin device (compare with the dirty bitset
 which is just for the smaller cache device).

 Target interface
 ================

 Constructor
 -----------

  cache <metadata dev> <cache dev> <origin dev> <block size>
        <#feature args> [<feature arg>]*
        <policy> <#policy args> [policy args]*

  metadata dev    : fast device holding the persistent metadata
  cache dev	 : fast device holding cached data blocks
  origin dev	 : slow device holding original data blocks
  block size      : cache unit size in sectors

  #feature args   : number of feature arguments passed
  feature args    : writethrough.  (The default is writeback.)

  policy          : the replacement policy to use
  #policy args    : an even number of arguments corresponding to
                    key/value pairs passed to the policy
  policy args     : key/value pairs passed to the policy
 		   E.g. 'sequential_threshold 1024'
 		   See cache-policies.txt for details.

 Optional feature arguments are:
    writethrough  : write through caching that prohibits cache block
 		   content from being different from origin block content.
 		   Without this argument, the default behaviour is to write
 		   back cache block contents later for performance reasons,
 		   so they may differ from the corresponding origin blocks.

 A policy called 'default' is always registered.  This is an alias for
 the policy we currently think is giving best all round performance.

 As the default policy could vary between kernels, if you are relying on
 the characteristics of a specific policy, always request it by name.

 Status
 ------

 <#used metadata blocks>/<#total metadata blocks> <#read hits> <#read misses>
 <#write hits> <#write misses> <#demotions> <#promotions> <#blocks in cache>
 <#dirty> <#features> <features>* <#core args> <core args>* <#policy args>
 <policy args>*

 #used metadata blocks    : Number of metadata blocks used
 #total metadata blocks   : Total number of metadata blocks
 #read hits               : Number of times a READ bio has been mapped
 			     to the cache
 #read misses             : Number of times a READ bio has been mapped
 			     to the origin
 #write hits              : Number of times a WRITE bio has been mapped
 			     to the cache
 #write misses            : Number of times a WRITE bio has been
 			     mapped to the origin
 #demotions               : Number of times a block has been removed
 			     from the cache
 #promotions              : Number of times a block has been moved to
 			     the cache
 #blocks in cache         : Number of blocks resident in the cache
 #dirty                   : Number of blocks in the cache that differ
 			     from the origin
 #feature args            : Number of feature args to follow
 feature args             : 'writethrough' (optional)
 #core args               : Number of core arguments (must be even)
 core args                : Key/value pairs for tuning the core
 			     e.g. migration_threshold
 #policy args             : Number of policy arguments to follow (must be even)
 policy args              : Key/value pairs
 			     e.g. 'sequential_threshold 1024

 Messages
 --------

 Policies will have different tunables, specific to each one, so we
 need a generic way of getting and setting these.  Device-mapper
 messages are used.  (A sysfs interface would also be possible.)

 The message format is:

    <key> <value>

 E.g.
    dmsetup message my_cache 0 sequential_threshold 1024

 Examples
 ========

 The test suite can be found here:

 https://github.com/jthornber/thinp-test-suite

 dmsetup create my_cache --table '0 41943040 cache /dev/mapper/metadata \
 	/dev/mapper/ssd /dev/mapper/origin 512 1 writeback default 0'
 dmsetup create my_cache --table '0 41943040 cache /dev/mapper/metadata \
 	/dev/mapper/ssd /dev/mapper/origin 1024 1 writeback \
 	mq 4 sequential_threshold 1024 random_threshold 8'
	Introduction
	============

	dm-cache is a device mapper target written by Joe Thornber, Heinz
	Mauelshagen, and Mike Snitzer.

	It aims to improve performance of a block device (eg, a spindle) by
	dynamically migrating some of its data to a faster, smaller device
	(eg, an SSD).

	This device-mapper solution allows us to insert this caching at
	different levels of the dm stack, for instance above the data device for
	a thin-provisioning pool. Caching solutions that are integrated more
	closely with the virtual memory system should give better performance.

	The target reuses the metadata library used in the thin-provisioning
	library.

	The decision as to what data to migrate and when is left to a plug-in
	policy module. Several of these have been written as we experiment,
	and we hope other people will contribute others for specific io
	scenarios (eg. a vm image server).

	Glossary
	========

	Migration - Movement of the primary copy of a logical block from one
	device to the other.
	Promotion - Migration from slow device to fast device.
	Demotion - Migration from fast device to slow device.

	The origin device always contains a copy of the logical block, which
	may be out of date or kept in sync with the copy on the cache device
	(depending on policy).

	Design
	======

	Sub-devices
	-----------

	The target is constructed by passing three devices to it (along with
	other parameters detailed later):

	1. An origin device - the big, slow one.

	2. A cache device - the small, fast one.

	3. A small metadata device - records which blocks are in the cache,
	which are dirty, and extra hints for use by the policy object.
	This information could be put on the cache device, but having it
	separate allows the volume manager to configure it differently,
	e.g. as a mirror for extra robustness.

	Fixed block size
	----------------

	The origin is divided up into blocks of a fixed size. This block size
	is configurable when you first create the cache. Typically we've been
	using block sizes of 256k - 1024k.

	Having a fixed block size simplifies the target a lot. But it is
	something of a compromise. For instance, a small part of a block may be
	getting hit a lot, yet the whole block will be promoted to the cache.
	So large block sizes are bad because they waste cache space. And small
	block sizes are bad because they increase the amount of metadata (both
	in core and on disk).

	Writeback/writethrough
	----------------------

	The cache has two modes, writeback and writethrough.

	If writeback, the default, is selected then a write to a block that is
	cached will go only to the cache and the block will be marked dirty in
	the metadata.

	If writethrough is selected then a write to a cached block will not
	complete until it has hit both the origin and cache devices. Clean
	blocks should remain clean.

	A simple cleaner policy is provided, which will clean (write back) all
	dirty blocks in a cache. Useful for decommissioning a cache.

	Migration throttling
	--------------------

	Migrating data between the origin and cache device uses bandwidth.
	The user can set a throttle to prevent more than a certain amount of
	migration occuring at any one time. Currently we're not taking any
	account of normal io traffic going to the devices. More work needs
	doing here to avoid migrating during those peak io moments.

	For the time being, a message "migration_threshold <#sectors>"
	can be used to set the maximum number of sectors being migrated,
	the default being 204800 sectors (or 100MB).

	Updating on-disk metadata
	-------------------------

	On-disk metadata is committed every time a REQ_SYNC or REQ_FUA bio is
	written. If no such requests are made then commits will occur every
	second. This means the cache behaves like a physical disk that has a
	write cache (the same is true of the thin-provisioning target). If
	power is lost you may lose some recent writes. The metadata should
	always be consistent in spite of any crash.

	The 'dirty' state for a cache block changes far too frequently for us
	to keep updating it on the fly. So we treat it as a hint. In normal
	operation it will be written when the dm device is suspended. If the
	system crashes all cache blocks will be assumed dirty when restarted.

	Per-block policy hints
	----------------------

	Policy plug-ins can store a chunk of data per cache block. It's up to
	the policy how big this chunk is, but it should be kept small. Like the
	dirty flags this data is lost if there's a crash so a safe fallback
	value should always be possible.

	For instance, the 'mq' policy, which is currently the default policy,
	uses this facility to store the hit count of the cache blocks. If
	there's a crash this information will be lost, which means the cache
	may be less efficient until those hit counts are regenerated.

	Policy hints affect performance, not correctness.

	Policy messaging
	----------------

	Policies will have different tunables, specific to each one, so we
	need a generic way of getting and setting these. Device-mapper
	messages are used. Refer to cache-policies.txt.

	Discard bitset resolution
	-------------------------

	We can avoid copying data during migration if we know the block has
	been discarded. A prime example of this is when mkfs discards the
	whole block device. We store a bitset tracking the discard state of
	blocks. However, we allow this bitset to have a different block size
	from the cache blocks. This is because we need to track the discard
	state for all of the origin device (compare with the dirty bitset
	which is just for the smaller cache device).

	Target interface
	================

	Constructor
	-----------

	cache <metadata dev> <cache dev> <origin dev> <block size>
	<#feature args> [<feature arg>]*
	<policy> <#policy args> [policy args]*

	metadata dev : fast device holding the persistent metadata
	cache dev : fast device holding cached data blocks
	origin dev : slow device holding original data blocks
	block size : cache unit size in sectors

	#feature args : number of feature arguments passed
	feature args : writethrough. (The default is writeback.)

	policy : the replacement policy to use
	#policy args : an even number of arguments corresponding to
	key/value pairs passed to the policy
	policy args : key/value pairs passed to the policy
	E.g. 'sequential_threshold 1024'
	See cache-policies.txt for details.

	Optional feature arguments are:
	writethrough : write through caching that prohibits cache block
	content from being different from origin block content.
	Without this argument, the default behaviour is to write
	back cache block contents later for performance reasons,
	so they may differ from the corresponding origin blocks.

	A policy called 'default' is always registered. This is an alias for
	the policy we currently think is giving best all round performance.

	As the default policy could vary between kernels, if you are relying on
	the characteristics of a specific policy, always request it by name.

	Status
	------

	<#used metadata blocks>/<#total metadata blocks> <#read hits> <#read misses>
	<#write hits> <#write misses> <#demotions> <#promotions> <#blocks in cache>
	<#dirty> <#features> <features>* <#core args> <core args>* <#policy args>
	<policy args>*

	#used metadata blocks : Number of metadata blocks used
	#total metadata blocks : Total number of metadata blocks
	#read hits : Number of times a READ bio has been mapped
	to the cache
	#read misses : Number of times a READ bio has been mapped
	to the origin
	#write hits : Number of times a WRITE bio has been mapped
	to the cache
	#write misses : Number of times a WRITE bio has been
	mapped to the origin
	#demotions : Number of times a block has been removed
	from the cache
	#promotions : Number of times a block has been moved to
	the cache
	#blocks in cache : Number of blocks resident in the cache
	#dirty : Number of blocks in the cache that differ
	from the origin
	#feature args : Number of feature args to follow
	feature args : 'writethrough' (optional)
	#core args : Number of core arguments (must be even)
	core args : Key/value pairs for tuning the core
	e.g. migration_threshold
	#policy args : Number of policy arguments to follow (must be even)
	policy args : Key/value pairs
	e.g. 'sequential_threshold 1024

	Messages
	--------

	Policies will have different tunables, specific to each one, so we
	need a generic way of getting and setting these. Device-mapper
	messages are used. (A sysfs interface would also be possible.)

	The message format is:

	<key> <value>

	E.g.
	dmsetup message my_cache 0 sequential_threshold 1024

	Examples
	========

	The test suite can be found here:

	https://github.com/jthornber/thinp-test-suite

	dmsetup create my_cache --table '0 41943040 cache /dev/mapper/metadata \
	/dev/mapper/ssd /dev/mapper/origin 512 1 writeback default 0'
	dmsetup create my_cache --table '0 41943040 cache /dev/mapper/metadata \
	/dev/mapper/ssd /dev/mapper/origin 1024 1 writeback \
	mq 4 sequential_threshold 1024 random_threshold 8'