summaryrefslogtreecommitdiffstats
path: root/doc/rados/operations/cache-tiering.rst
blob: 8056ace4743bd628a8b6296c693bb08bd3230a2e (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
===============
 Cache Tiering
===============

A cache tier provides Ceph Clients with better I/O performance for a subset of
the data stored in a backing storage tier. Cache tiering involves creating a
pool of relatively fast/expensive storage devices (e.g., solid state drives)
configured to act as a cache tier, and a backing pool of either erasure-coded
or relatively slower/cheaper devices configured to act as an economical storage
tier. The Ceph objecter handles where to place the objects and the tiering
agent determines when to flush objects from the cache to the backing storage
tier. So the cache tier and the backing storage tier are completely transparent 
to Ceph clients.


.. ditaa::
           +-------------+
           | Ceph Client |
           +------+------+
                  ^
     Tiering is   |  
    Transparent   |              Faster I/O
        to Ceph   |           +---------------+
     Client Ops   |           |               |   
                  |    +----->+   Cache Tier  |
                  |    |      |               |
                  |    |      +-----+---+-----+
                  |    |            |   ^ 
                  v    v            |   |   Active Data in Cache Tier
           +------+----+--+         |   |
           |   Objecter   |         |   |
           +-----------+--+         |   |
                       ^            |   |   Inactive Data in Storage Tier
                       |            v   |
                       |      +-----+---+-----+
                       |      |               |
                       +----->|  Storage Tier |
                              |               |
                              +---------------+
                                 Slower I/O


The cache tiering agent handles the migration of data between the cache tier 
and the backing storage tier automatically. However, admins have the ability to
configure how this migration takes place by setting the ``cache-mode``. There are
two main scenarios:

- **writeback** mode: If the base tier and the cache tier are configured in
  ``writeback`` mode, Ceph clients receive an ACK from the base tier every time
  they write data to it. Then the cache tiering agent determines whether
  ``osd_tier_default_cache_min_write_recency_for_promote`` has been set. If it
  has been set and the data has been written more than a specified number of
  times per interval, the data is promoted to the cache tier.

  When Ceph clients need access to data stored in the base tier, the cache
  tiering agent reads the data from the base tier and returns it to the client.
  While data is being read from the base tier, the cache tiering agent consults
  the value of ``osd_tier_default_cache_min_read_recency_for_promote`` and
  decides whether to promote that data from the base tier to the cache tier.
  When data has been promoted from the base tier to the cache tier, the Ceph
  client is able to perform I/O operations on it using the cache tier. This is
  well-suited for mutable data (for example, photo/video editing, transactional
  data).

- **readproxy** mode: This mode will use any objects that already
  exist in the cache tier, but if an object is not present in the
  cache the request will be proxied to the base tier.  This is useful
  for transitioning from ``writeback`` mode to a disabled cache as it
  allows the workload to function properly while the cache is drained,
  without adding any new objects to the cache.

Other cache modes are:

- **readonly** promotes objects to the cache on read operations only; write
  operations are forwarded to the base tier. This mode is intended for
  read-only workloads that do not require consistency to be enforced by the
  storage system. (**Warning**: when objects are updated in the base tier,
  Ceph makes **no** attempt to sync these updates to the corresponding objects
  in the cache. Since this mode is considered experimental, a
  ``--yes-i-really-mean-it`` option must be passed in order to enable it.)

- **none** is used to completely disable caching.


A word of caution
=================

Cache tiering will *degrade* performance for most workloads.  Users should use
extreme caution before using this feature.

* *Workload dependent*: Whether a cache will improve performance is
  highly dependent on the workload.  Because there is a cost
  associated with moving objects into or out of the cache, it can only
  be effective when there is a *large skew* in the access pattern in
  the data set, such that most of the requests touch a small number of
  objects.  The cache pool should be large enough to capture the
  working set for your workload to avoid thrashing.

* *Difficult to benchmark*: Most benchmarks that users run to measure
  performance will show terrible performance with cache tiering, in
  part because very few of them skew requests toward a small set of
  objects, it can take a long time for the cache to "warm up," and
  because the warm-up cost can be high.

* *Usually slower*: For workloads that are not cache tiering-friendly,
  performance is often slower than a normal RADOS pool without cache
  tiering enabled.

* *librados object enumeration*: The librados-level object enumeration
  API is not meant to be coherent in the presence of the case.  If
  your application is using librados directly and relies on object
  enumeration, cache tiering will probably not work as expected.
  (This is not a problem for RGW, RBD, or CephFS.)

* *Complexity*: Enabling cache tiering means that a lot of additional
  machinery and complexity within the RADOS cluster is being used.
  This increases the probability that you will encounter a bug in the system
  that other users have not yet encountered and will put your deployment at a
  higher level of risk.

Known Good Workloads
--------------------

* *RGW time-skewed*: If the RGW workload is such that almost all read
  operations are directed at recently written objects, a simple cache
  tiering configuration that destages recently written objects from
  the cache to the base tier after a configurable period can work
  well.

Known Bad Workloads
-------------------

The following configurations are *known to work poorly* with cache
tiering.

* *RBD with replicated cache and erasure-coded base*: This is a common
  request, but usually does not perform well.  Even reasonably skewed
  workloads still send some small writes to cold objects, and because
  small writes are not yet supported by the erasure-coded pool, entire
  (usually 4 MB) objects must be migrated into the cache in order to
  satisfy a small (often 4 KB) write.  Only a handful of users have
  successfully deployed this configuration, and it only works for them
  because their data is extremely cold (backups) and they are not in
  any way sensitive to performance.

* *RBD with replicated cache and base*: RBD with a replicated base
  tier does better than when the base is erasure coded, but it is
  still highly dependent on the amount of skew in the workload, and
  very difficult to validate.  The user will need to have a good
  understanding of their workload and will need to tune the cache
  tiering parameters carefully.


Setting Up Pools
================

To set up cache tiering, you must have two pools. One will act as the 
backing storage and the other will act as the cache.


Setting Up a Backing Storage Pool
---------------------------------

Setting up a backing storage pool typically involves one of two scenarios: 

- **Standard Storage**: In this scenario, the pool stores multiple copies
  of an object in the Ceph Storage Cluster.

- **Erasure Coding:** In this scenario, the pool uses erasure coding to 
  store data much more efficiently with a small performance tradeoff.

In the standard storage scenario, you can setup a CRUSH rule to establish 
the failure domain (e.g., osd, host, chassis, rack, row, etc.). Ceph OSD 
Daemons perform optimally when all storage drives in the rule are of the 
same size, speed (both RPMs and throughput) and type. See `CRUSH Maps`_ 
for details on creating a rule. Once you have created a rule, create 
a backing storage pool. 

In the erasure coding scenario, the pool creation arguments will generate the
appropriate rule automatically. See `Create a Pool`_ for details.

In subsequent examples, we will refer to the backing storage pool 
as ``cold-storage``.


Setting Up a Cache Pool
-----------------------

Setting up a cache pool follows the same procedure as the standard storage
scenario, but with this difference: the drives for the cache tier are typically
high performance drives that reside in their own servers and have their own
CRUSH rule.  When setting up such a rule, it should take account of the hosts
that have the high performance drives while omitting the hosts that don't. See
:ref:`CRUSH Device Class<crush-map-device-class>` for details.


In subsequent examples, we will refer to the cache pool as ``hot-storage`` and
the backing pool as ``cold-storage``.

For cache tier configuration and default values, see 
`Pools - Set Pool Values`_.


Creating a Cache Tier
=====================

Setting up a cache tier involves associating a backing storage pool with
a cache pool:

.. prompt:: bash $

   ceph osd tier add {storagepool} {cachepool}

For example:

.. prompt:: bash $

   ceph osd tier add cold-storage hot-storage

To set the cache mode, execute the following:

.. prompt:: bash $

   ceph osd tier cache-mode {cachepool} {cache-mode}

For example:

.. prompt:: bash $

   ceph osd tier cache-mode hot-storage writeback

The cache tiers overlay the backing storage tier, so they require one
additional step: you must direct all client traffic from the storage pool to 
the cache pool. To direct client traffic directly to the cache pool, execute 
the following:

.. prompt:: bash $

   ceph osd tier set-overlay {storagepool} {cachepool}

For example:

.. prompt:: bash $

   ceph osd tier set-overlay cold-storage hot-storage


Configuring a Cache Tier
========================

Cache tiers have several configuration options. You may set
cache tier configuration options with the following usage:

.. prompt:: bash $

   ceph osd pool set {cachepool} {key} {value}
   
See `Pools - Set Pool Values`_ for details.


Target Size and Type
--------------------

Ceph's production cache tiers use a `Bloom Filter`_ for the ``hit_set_type``:

.. prompt:: bash $

   ceph osd pool set {cachepool} hit_set_type bloom

For example:

.. prompt:: bash $

   ceph osd pool set hot-storage hit_set_type bloom

The ``hit_set_count`` and ``hit_set_period`` define how many such HitSets to
store, and how much time each HitSet should cover:

.. prompt:: bash $

   ceph osd pool set {cachepool} hit_set_count 12
   ceph osd pool set {cachepool} hit_set_period 14400
   ceph osd pool set {cachepool} target_max_bytes 1000000000000

.. note:: A larger ``hit_set_count`` results in more RAM consumed by
          the ``ceph-osd`` process.

Binning accesses over time allows Ceph to determine whether a Ceph client
accessed an object at least once, or more than once over a time period 
("age" vs "temperature").

The ``min_read_recency_for_promote`` defines how many HitSets to check for the
existence of an object when handling a read operation. The checking result is
used to decide whether to promote the object asynchronously. Its value should be
between 0 and ``hit_set_count``. If it's set to 0, the object is always promoted.
If it's set to 1, the current HitSet is checked. And if this object is in the
current HitSet, it's promoted. Otherwise not. For the other values, the exact
number of archive HitSets are checked. The object is promoted if the object is
found in any of the most recent ``min_read_recency_for_promote`` HitSets.

A similar parameter can be set for the write operation, which is
``min_write_recency_for_promote``:

.. prompt:: bash $

   ceph osd pool set {cachepool} min_read_recency_for_promote 2
   ceph osd pool set {cachepool} min_write_recency_for_promote 2

.. note:: The longer the period and the higher the
   ``min_read_recency_for_promote`` and
   ``min_write_recency_for_promote``values, the more RAM the ``ceph-osd``
   daemon consumes. In particular, when the agent is active to flush
   or evict cache objects, all ``hit_set_count`` HitSets are loaded
   into RAM.


Cache Sizing
------------

The cache tiering agent performs two main functions: 

- **Flushing:** The agent identifies modified (or dirty) objects and forwards
  them to the storage pool for long-term storage.
  
- **Evicting:** The agent identifies objects that haven't been modified 
  (or clean) and evicts the least recently used among them from the cache.


Absolute Sizing
~~~~~~~~~~~~~~~

The cache tiering agent can flush or evict objects based upon the total number
of bytes or the total number of objects. To specify a maximum number of bytes,
execute the following:

.. prompt:: bash $

   ceph osd pool set {cachepool} target_max_bytes {#bytes}

For example, to flush or evict at 1 TB, execute the following:

.. prompt:: bash $

   ceph osd pool set hot-storage target_max_bytes 1099511627776

To specify the maximum number of objects, execute the following:

.. prompt:: bash $

   ceph osd pool set {cachepool} target_max_objects {#objects}

For example, to flush or evict at 1M objects, execute the following:

.. prompt:: bash $

   ceph osd pool set hot-storage target_max_objects 1000000

.. note:: Ceph is not able to determine the size of a cache pool automatically, so
   the configuration on the absolute size is required here, otherwise the
   flush/evict will not work. If you specify both limits, the cache tiering
   agent will begin flushing or evicting when either threshold is triggered.

.. note:: All client requests will be blocked only when  ``target_max_bytes`` or
   ``target_max_objects`` reached

Relative Sizing
~~~~~~~~~~~~~~~

The cache tiering agent can flush or evict objects relative to the size of the
cache pool(specified by ``target_max_bytes`` / ``target_max_objects`` in
`Absolute sizing`_).  When the cache pool consists of a certain percentage of
modified (or dirty) objects, the cache tiering agent will flush them to the
storage pool. To set the ``cache_target_dirty_ratio``, execute the following:

.. prompt:: bash $

   ceph osd pool set {cachepool} cache_target_dirty_ratio {0.0..1.0}

For example, setting the value to ``0.4`` will begin flushing modified
(dirty) objects when they reach 40% of the cache pool's capacity:

.. prompt:: bash $

   ceph osd pool set hot-storage cache_target_dirty_ratio 0.4

When the dirty objects reaches a certain percentage of its capacity, flush dirty
objects with a higher speed. To set the ``cache_target_dirty_high_ratio``:

.. prompt:: bash $

   ceph osd pool set {cachepool} cache_target_dirty_high_ratio {0.0..1.0}

For example, setting the value to ``0.6`` will begin aggressively flush dirty
objects when they reach 60% of the cache pool's capacity. obviously, we'd
better set the value between dirty_ratio and full_ratio:

.. prompt:: bash $

   ceph osd pool set hot-storage cache_target_dirty_high_ratio 0.6

When the cache pool reaches a certain percentage of its capacity, the cache
tiering agent will evict objects to maintain free capacity. To set the 
``cache_target_full_ratio``, execute the following:

.. prompt:: bash $

   ceph osd pool set {cachepool} cache_target_full_ratio {0.0..1.0}

For example, setting the value to ``0.8`` will begin flushing unmodified
(clean) objects when they reach 80% of the cache pool's capacity:

.. prompt:: bash $

   ceph osd pool set hot-storage cache_target_full_ratio 0.8


Cache Age
---------

You can specify the minimum age of an object before the cache tiering agent 
flushes a recently modified (or dirty) object to the backing storage pool:

.. prompt:: bash $

   ceph osd pool set {cachepool} cache_min_flush_age {#seconds}

For example, to flush modified (or dirty) objects after 10 minutes, execute the
following:

.. prompt:: bash $

   ceph osd pool set hot-storage cache_min_flush_age 600

You can specify the minimum age of an object before it will be evicted from the
cache tier:

.. prompt:: bash $

   ceph osd pool {cache-tier} cache_min_evict_age {#seconds}

For example, to evict objects after 30 minutes, execute the following:

.. prompt:: bash $

   ceph osd pool set hot-storage cache_min_evict_age 1800


Removing a Cache Tier
=====================

Removing a cache tier differs depending on whether it is a writeback 
cache or a read-only cache.


Removing a Read-Only Cache
--------------------------

Since a read-only cache does not have modified data, you can disable
and remove it without losing any recent changes to objects in the cache. 

#. Change the cache-mode to ``none`` to disable it.:

   .. prompt:: bash 

      ceph osd tier cache-mode {cachepool} none

   For example:

   .. prompt:: bash $

      ceph osd tier cache-mode hot-storage none

#. Remove the cache pool from the backing pool.:

   .. prompt:: bash $

      ceph osd tier remove {storagepool} {cachepool}

   For example:

   .. prompt:: bash $

      ceph osd tier remove cold-storage hot-storage


Removing a Writeback Cache
--------------------------

Since a writeback cache may have modified data, you must take steps to ensure 
that you do not lose any recent changes to objects in the cache before you 
disable and remove it.


#. Change the cache mode to ``proxy`` so that new and modified objects will 
   flush to the backing storage pool.:

   .. prompt:: bash $

      ceph osd tier cache-mode {cachepool} proxy

   For example: 

   .. prompt:: bash $

      ceph osd tier cache-mode hot-storage proxy


#. Ensure that the cache pool has been flushed. This may take a few minutes:

   .. prompt:: bash $

      rados -p {cachepool} ls

   If the cache pool still has objects, you can flush them manually. 
   For example:

   .. prompt:: bash $

      rados -p {cachepool} cache-flush-evict-all


#. Remove the overlay so that clients will not direct traffic to the cache.:

   .. prompt:: bash $

      ceph osd tier remove-overlay {storagetier}

   For example:

   .. prompt:: bash $

      ceph osd tier remove-overlay cold-storage


#. Finally, remove the cache tier pool from the backing storage pool.:

   .. prompt:: bash $

      ceph osd tier remove {storagepool} {cachepool} 

   For example:

   .. prompt:: bash $

      ceph osd tier remove cold-storage hot-storage


.. _Create a Pool: ../pools#create-a-pool
.. _Pools - Set Pool Values: ../pools#set-pool-values
.. _Bloom Filter: https://en.wikipedia.org/wiki/Bloom_filter
.. _CRUSH Maps: ../crush-map
.. _Absolute Sizing: #absolute-sizing