1VMOD(DIRECTORS) VMOD(DIRECTORS)
2
6 VMOD directors - Varnish Directors Module
7
9 import directors [as name] [from "path"]
10 new xround_robin = directors.round_robin()
12 VOID xround_robin.add_backend(BACKEND)
14 VOID xround_robin.remove_backend(BACKEND)
16 BACKEND xround_robin.backend()
18 new xfallback = directors.fallback(BOOL sticky)
20 VOID xfallback.add_backend(BACKEND)
22 VOID xfallback.remove_backend(BACKEND)
24 BACKEND xfallback.backend()
26 new xrandom = directors.random()
28 VOID xrandom.add_backend(BACKEND, REAL)
30 VOID xrandom.remove_backend(BACKEND)
32 BACKEND xrandom.backend()
34 new xhash = directors.hash()
36 VOID xhash.add_backend(BACKEND, REAL)
38 VOID xhash.remove_backend(BACKEND)
40 BACKEND xhash.backend(STRING)
42 new xshard = directors.shard()
44 VOID xshard.set_warmup(REAL probability)
46 VOID xshard.set_rampup(DURATION duration)
48 VOID xshard.associate(BLOB param)
50 BOOL xshard.add_backend(BACKEND backend, [STRING ident], [DURATION rampup])
52 BOOL xshard.remove_backend([BACKEND backend], [STRING ident])
54 BOOL xshard.clear()
56 BOOL xshard.reconfigure(INT replicas)
58 INT xshard.key(STRING)
60 BACKEND xshard.backend([ENUM by], [INT key], [BLOB key_blob], [INT alt], [REAL warmup], [BOOL rampup], [ENUM healthy], [BLOB param], [ENUM resolve])
62 VOID xshard.debug(INT)
64 new xshard_param = directors.shard_param()
66 VOID xshard_param.clear()
68 VOID xshard_param.set([ENUM by], [INT key], [BLOB key_blob], [INT alt], [REAL warmup], [BOOL rampup], [ENUM healthy])
70 STRING xshard_param.get_by()
72 INT xshard_param.get_key()
74 INT xshard_param.get_alt()
76 REAL xshard_param.get_warmup()
78 BOOL xshard_param.get_rampup()
80 STRING xshard_param.get_healthy()
82 BLOB xshard_param.use()
84 BACKEND lookup(STRING)
86
88 vmod_directors enables backend load balancing in Varnish.
89 The module implements load balancing techniques, and also serves as an
91 example on how one could extend the load balancing capabilities of Var‐
92 nish.
93 To enable load balancing you must import this vmod (directors).
95 Then you define your backends. Once you have the backends declared you
97 can add them to a director. This happens in executed VCL code. If you
98 want to emulate the previous behavior of Varnish 3.0 you can just ini‐
99 tialize the directors in vcl_init{}, like this:
100 sub vcl_init {
102 new vdir = directors.round_robin();
103 vdir.add_backend(backend1);
104 vdir.add_backend(backend2);
105 }
106 As you can see there is nothing keeping you from manipulating the
108 directors elsewhere in VCL. So, you could have VCL code that would add
109 more backends to a director when a certain URL is called.
110 Note that directors can use other directors as backends.
112 new xround_robin = directors.round_robin()
114 Create a round robin director.
115 This director will pick backends in a round robin fashion.
117 Example:
119 new vdir = directors.round_robin();
121 VOID xround_robin.add_backend(BACKEND)
123 Add a backend to the round-robin director.
124 Example:
126 vdir.add_backend(backend1);
128 VOID xround_robin.remove_backend(BACKEND)
130 Remove a backend from the round-robin director.
131 Example:
133 vdir.remove_backend(backend1);
135 BACKEND xround_robin.backend()
137 Pick a backend from the director.
138 Example:
140 set req.backend_hint = vdir.backend();
142 new xfallback = directors.fallback(BOOL sticky=0)
144 Create a fallback director.
145 A fallback director will try each of the added backends in turn, and
147 return the first one that is healthy.
148 If sticky is set to true, the director will keep using the healthy
150 backend, even if a higher-priority backend becomes available. Once the
151 whole backend list is exhausted, it'll start over at the beginning.
152 Example:
154 new vdir = directors.fallback();
156 VOID xfallback.add_backend(BACKEND)
158 Add a backend to the director.
159 Note that the order in which this is done matters for the fallback
161 director.
162 Example:
164 vdir.add_backend(backend1);
166 VOID xfallback.remove_backend(BACKEND)
168 Remove a backend from the director.
169 Example:
171 vdir.remove_backend(backend1);
173 BACKEND xfallback.backend()
175 Pick a backend from the director.
176 Example:
178 set req.backend_hint = vdir.backend();
180 new xrandom = directors.random()
182 Create a random backend director.
183 The random director distributes load over the backends using a weighted
185 random probability distribution.
186 The "testable" random generator in varnishd is used, which enables
188 deterministic tests to be run (See: d00004.vtc).
189 Example:
191 new vdir = directors.random();
193 VOID xrandom.add_backend(BACKEND, REAL)
195 Add a backend to the director with a given weight.
196 Each backend will receive approximately 100 * (weight /
198 (sum(all_added_weights))) per cent of the traffic sent to this direc‐
199 tor.
200 Example:
202 # 2/3 to backend1, 1/3 to backend2.
204 vdir.add_backend(backend1, 10.0);
205 vdir.add_backend(backend2, 5.0);
206 VOID xrandom.remove_backend(BACKEND)
208 Remove a backend from the director.
209 Example:
211 vdir.remove_backend(backend1);
213 BACKEND xrandom.backend()
215 Pick a backend from the director.
216 Example:
218 set req.backend_hint = vdir.backend();
220 new xhash = directors.hash()
222 Create a hashing backend director.
223 The director chooses the backend server by computing a hash/digest of
225 the string given to xhash.backend().
226 Commonly used with client.ip or a session cookie to get sticky ses‐
228 sions.
229 Example:
231 new vdir = directors.hash();
233 VOID xhash.add_backend(BACKEND, REAL)
235 Add a backend to the director with a certain weight.
236 Weight is used as in the random director. Recommended value is 1.0
238 unless you have special needs.
239 Example:
241 vdir.add_backend(backend1, 1.0);
243 VOID xhash.remove_backend(BACKEND)
245 Remove a backend from the director.
246 Example::
248 vdir.remove_backend(backend1);
249 BACKEND xhash.backend(STRING)
251 Pick a backend from the backend director.
252 Use the string or list of strings provided to pick the backend.
254 Example::
256 # pick a backend based on the cookie header from the client set
257 req.backend_hint = vdir.backend(req.http.cookie);
258 new xshard = directors.shard()
260 Create a shard director.
261 Note that the shard director needs to be configured using at least one
263 xshard.add_backend() call(s) followed by a xshard.reconfigure() call
264 before it can hand out backends.
265 _Note_ that due to various restrictions (documented below), it is rec‐
267 ommended to use the shard director on the backend side.
268 Introduction
270 The shard director selects backends by a key, which can be provided
271 directly or derived from strings. For the same key, the shard director
272 will always return the same backend, unless the backend configuration
273 or health state changes. Conversely, for differing keys, the shard
274 director will likely choose different backends. In the default configu‐
275 ration, unhealthy backends are not selected.
276 The shard director resembles the hash director, but its main advantage
278 is that, when the backend configuration or health states change, the
279 association of keys to backends remains as stable as possible.
280 In addition, the rampup and warmup features can help to further improve
282 user-perceived response times.
283 Sharding
285 This basic technique allows for numerous applications like optimizing
286 backend server cache efficiency, Varnish clustering or persisting ses‐
287 sions to servers without keeping any state, and, in particular, without
288 the need to synchronize state between nodes of a cluster of Varnish
289 servers:
290 · Many applications use caches for data objects, so, in a cluster of
292 application servers, requesting similar objects from the same server
293 may help to optimize efficiency of such caches.
294 For example, sharding by URL or some id component of the url has been
296 shown to drastically improve the efficiency of many content manage‐
297 ment systems.
298 · As special case of the previous example, in clusters of Varnish
300 servers without additional request distribution logic, each cache
301 will need store all hot objects, so the effective cache size is
302 approximately the smallest cache size of any server in the cluster.
303 Sharding allows to segregate objects within the cluster such that
305 each object is only cached on one of the servers (or on one primary
306 and one backup, on a primary for long and others for short etc...).
307 Effectively, this will lead to a cache size in the order of the sum
308 of all individual caches, with the potential to drastically increase
309 efficiency (scales by the number of servers).
310 · Another application is to implement persistence of backend requests,
312 such that all requests sharing a certain criterion (such as an IP
313 address or session ID) get forwarded to the same backend server.
314 When used with clusters of varnish servers, the shard director will, if
316 otherwise configured equally, make the same decision on all servers. In
317 other words, requests sharing a common criterion used as the shard key
318 will be balanced onto the same backend server(s) no matter which Var‐
319 nish server handles the request.
320 The drawbacks are:
322 · the distribution of requests depends on the number of requests per
324 key and the uniformity of the distribution of key values. In short,
325 while this technique may lead to much better efficiency overall, it
326 may also lead to less good load balancing for specific cases.
327 · When a backend server becomes unavailable, every persistence tech‐
329 nique has to reselect a new backend server, but this technique will
330 also switch back to the preferred server once it becomes healthy
331 again, so when used for persistence, it is generally less stable com‐
332 pared to stateful techniques (which would continue to use a selected
333 server for as long as possible (or dictated by a TTL)).
334 Method
336 When xshard.reconfigure() is called, a consistent hashing circular data
337 structure gets built from the last 32 bits of SHA256 hash values of
338 <ident><n> (default ident being the backend name) for each backend and
339 for a running number n from 1 to replicas. Hashing creates the seem‐
340 ingly random order for placement of backends on the consistent hashing
341 ring.
342 When xshard.backend() is called, a load balancing key gets generated
344 unless provided. The smallest hash value in the circle is looked up
345 that is larger than the key (searching clockwise and wrapping around as
346 necessary). The backend for this hash value is the preferred backend
347 for the given key.
348 If a healthy backend is requested, the search is continued linearly on
350 the ring as long as backends found are unhealthy or all backends have
351 been checked. The order of these "alternative backends" on the ring is
352 likely to differ for different keys. Alternative backends can also be
353 selected explicitly.
354 On consistent hashing see:
356 · http://www8.org/w8-papers/2a-webserver/caching/paper2.html
358 · http://www.audioscrobbler.net/development/ketama/
360 · svn://svn.audioscrobbler.net/misc/ketama
362 · http://en.wikipedia.org/wiki/Consistent_hashing
364 Error Reporting
366 Failing methods should report errors to VSL with the Error tag, so when
367 configuring the shard director, you are advised to check:
368 varnishlog -I Error:^shard
370 VOID xshard.set_warmup(REAL probability=0.0)
372 Set the default warmup probability. See the warmup parameter of
373 xshard.backend(). If probability is 0.0 (default), warmup is disabled.
374 VOID xshard.set_rampup(DURATION duration=0)
376 Set the default rampup duration. See rampup parameter of
377 xshard.backend(). If duration is 0 (default), rampup is disabled.
378 VOID xshard.associate(BLOB param=0)
380 Associate a default directors.shard_param() object or clear an associa‐
381 tion.
382 The value of the param argument must be a call to the
384 xshard_param.use() method. No argument clears the association.
385 The association can be changed per backend request using the param
387 argument of xshard.backend().
388 BOOL xshard.add_backend(BACKEND backend, [STRING ident], [DURATION rampup])
390 BOOL xshard.add_backend(
391 BACKEND backend,
392 [STRING ident],
393 [DURATION rampup]
394 )
395 Add a backend backend to the director.
397 ident: Optionally specify an identification string for this backend,
399 which will be hashed by xshard.reconfigure() to construct the consis‐
400 tent hashing ring. The identification string defaults to the backend
401 name.
402 ident allows to add multiple instances of the same backend.
404 rampup: Optionally specify a specific rampup time for this backend.
406 Otherwise, the per-director rampup time is used (see
407 xshard.set_rampup()).
408 NOTE: Backend changes need to be finalized with xshard.reconfigure()
410 and are only supported on one shard director at a time.
411 BOOL xshard.remove_backend([BACKEND backend], [STRING ident])
413 BOOL xshard.remove_backend(
414 [BACKEND backend=0],
415 [STRING ident=0]
416 )
417 Remove backend(s) from the director. Either backend or ident must be
419 specified. ident removes a specific instance. If backend is given with‐
420 out ident, all instances of this backend are removed.
421 NOTE: Backend changes need to be finalized with xshard.reconfigure()
423 and are only supported on one shard director at a time.
424 BOOL xshard.clear()
426 Remove all backends from the director.
427 NOTE: Backend changes need to be finalized with xshard.reconfigure()
429 and are only supported on one shard director at a time.
430 BOOL xshard.reconfigure(INT replicas=67)
432 Reconfigure the consistent hashing ring to reflect backend changes.
433 This method must be called at least once before the director can be
435 used.
436 INT xshard.key(STRING)
438 Convenience method to generate a sharding key for use with the key
439 argument to the xshard.backend() method by hashing the given string
440 with SHA256.
441 To generate sharding keys using other hashes, use a custom vmod like
443 vmod blobdigest with the key_blob argument of the xshard.backend()
444 method.
445 BACKEND xshard.backend([ENUM by], [INT key], [BLOB key_blob], [INT alt],
447 [REAL warmup], [BOOL rampup], [ENUM healthy], [BLOB param], [ENUM
448 resolve])
449 BACKEND xshard.backend(
450 [ENUM {HASH, URL, KEY, BLOB} by=HASH],
451 [INT key],
452 [BLOB key_blob],
453 [INT alt=0],
454 [REAL warmup=-1],
455 [BOOL rampup=1],
456 [ENUM {CHOSEN, IGNORE, ALL} healthy=CHOSEN],
457 [BLOB param],
458 [ENUM {NOW, LAZY} resolve]
459 )
460 Lookup a backend on the consistent hashing ring.
462 This documentation uses the notion of an order of backends for a par‐
464 ticular shard key. This order is deterministic but seemingly random as
465 determined by the consistent hashing algorithm and is likely to differ
466 for different keys, depending on the number of backends and the number
467 of replicas. In particular, the backend order referred to here is _not_
468 the order given when backends are added.
469 · by how to determine the sharding key
471 · HASH:
473 · when called in backend context: Use the varnish hash value as set
475 by vcl_hash{}
476 · when called in client context: hash req.url
478 · URL: hash req.url / bereq.url
480 · KEY: use the key argument
482 · BLOB: use the key_blob argument
484 · key lookup key with by=KEY
486 the xshard.key() method may come handy to generate a sharding key
488 from custom strings.
489 · key_blob lookup key with by=BLOB
491 Currently, this uses the first 4 bytes from the given blob in network
493 byte order (big endian), left-padded with zeros for blobs smaller
494 than 4 bytes.
495 · alt alternative backend selection
497 Select the alt-th alternative backend for the given key.
499 This is particularly useful for retries / restarts due to backend
501 errors: By setting alt=req.restarts or alt=bereq.retries with
502 healthy=ALL, another server gets selected.
503 The rampup and warmup features are only active for alt==0
505 · rampup slow start for servers which just went healthy
507 If alt==0 and the chosen backend is in its rampup period, with a
509 probability proportional to the fraction of time since the backup
510 became healthy to the rampup period, return the next alternative
511 backend, unless this is also in its rampup period.
512 The default rampup interval can be set per shard director using the
514 xshard.set_rampup() method or specifically per backend with the
515 xshard.add_backend() method.
516 · warmup probabilistic alternative server selection
518 possible values: -1, 0..1
520 -1: use the warmup probability from the director definition
522 Only used for alt==0: Sets the ratio of requests (0.0 to 1.0) that
524 goes to the next alternate backend to warm it up when the preferred
525 backend is healthy. Not active if any of the preferred or alternative
526 backend are in rampup.
527 warmup=0.5 is a convenient way to spread the load for each key over
529 two backends under normal operating conditions.
530 · healthy
532 · CHOSEN: Return a healthy backend if possible.
534 For alt==0, return the first healthy backend or none.
536 For alt > 0, ignore the health state of backends skipped for alter‐
538 native backend selection, then return the next healthy backend. If
539 this does not exist, return the last healthy backend of those
540 skipped or none.
541 · IGNORE: Completely ignore backend health state
543 Just return the first or alt-th alternative backend, ignoring
545 health state, rampup and warmup.
546 · ALL: Check health state also for alternative backend selection
548 For alt > 0, return the alt-th alternative backend of all those
550 healthy, the last healthy backend found or none.
551 · resolve
553 default: LAZY in vcl_init{}, NOW otherwise
555 · NOW: look up a backend and return it.
557 Can not be used in vcl_init{}.
559 · LAZY: return an instance of this director for later backend resolu‐
561 tion.
562 LAZY mode is required for referencing shard director instances, for
564 example as backends for other directors (director layering).
565 In vcl_init{} and on the client side, LAZY mode can not be used
567 with any other argument.
568 On the backend side, parameters from arguments or an associated
570 parameter set affect the shard director instance for the backend
571 request irrespective of where it is referenced.
572 · param
574 Use or associate a parameter set. The value of the param argument
576 must be a call to the xshard_param.use() method.
577 default: as set by xshard.associate() or unset.
579 · for resolve=NOW take parameter defaults from the
581 directors.shard_param() parameter set
582 · for resolve=LAZY associate the directors.shard_param() parameter
584 set for this backend request
585 Implementation notes for use of parameter sets with resolve=LAZY:
587 · A param argument remains associated and any changes to the asso‐
589 ciated parameter set affect the sharding decision once the direc‐
590 tor resolves to an actual backend.
591 · If other parameter arguments are also given, they have preference
593 and are kept even if the parameter set given by the param argu‐
594 ment is subsequently changed within the same backend request.
595 · Each call to xshard.backend() overrides any previous call.
597 VOID xshard.debug(INT)
599 intentionally undocumented
600 new xshard_param = directors.shard_param()
602 Create a shard parameter set.
603 A parameter set allows for re-use of xshard.backend() arguments across
605 many shard director instances and simplifies advanced use cases (e.g.
606 shard director with custom parameters layered below other directors).
607 Parameter sets have two scopes:
609 · per-VCL scope defined in vcl_init{}
611 · per backend request scope
613 The per-VCL scope defines defaults for the per backend scope. Any
615 changes to a parameter set in backend context only affect the respec‐
616 tive backend request.
617 Parameter sets can not be used in client context.
619 The following example is a typical use case: A parameter set is associ‐
621 ated with several directors. Director choice happens on the client side
622 and parameters are changed on the backend side to implement retries on
623 alternative backends:
624 sub vcl_init {
626 new shard_param = directors.shard_param();
627 new dir_A = directors.shard();
629 dir_A.add_backend(...);
630 dir_A.reconfigure(shard_param);
631 dir_A.associate(shard_param.use()); # <-- !
632 new dir_B = directors.shard();
634 dir_B.add_backend(...);
635 dir_B.reconfigure(shard_param);
636 dir_B.associate(shard_param.use()); # <-- !
637 }
638 sub vcl_recv {
640 if (...) {
641 set req.backend_hint = dir_A.backend(resolve=LAZY);
642 } else {
643 set req.backend_hint = dir_B.backend(resolve=LAZY);
644 }
645 }
646 sub vcl_backend_fetch {
648 # changes dir_A and dir_B behaviour
649 shard_param.set(alt=bereq.retries);
650 }
651 VOID xshard_param.clear()
653 Reset the parameter set to default values as documented for
654 xshard.backend().
655 · in vcl_init{}, resets the parameter set default for this VCL
657 · in backend context, resets the parameter set for this backend request
659 to the VCL defaults
660 This method may not be used in client context
662 VOID xshard_param.set([ENUM by], [INT key], [BLOB key_blob], [INT alt],
664 [REAL warmup], [BOOL rampup], [ENUM healthy])
665 VOID xshard_param.set(
666 [ENUM {HASH, URL, KEY, BLOB} by],
667 [INT key],
668 [BLOB key_blob],
669 [INT alt],
670 [REAL warmup],
671 [BOOL rampup],
672 [ENUM {CHOSEN, IGNORE, ALL} healthy]
673 )
674 Change the given parameters of a parameter set as documented for
676 xshard.backend().
677 · in vcl_init{}, changes the parameter set default for this VCL
679 · in backend context, changes the parameter set for this backend
681 request, keeping the defaults set for this VCL for unspecified argu‐
682 ments.
683 This method may not be used in client context
685 STRING xshard_param.get_by()
687 Get a string representation of the by enum argument which denotes how a
688 shard director using this parameter object would derive the shard key.
689 See xshard.backend().
690 INT xshard_param.get_key()
692 Get the key which a shard director using this parameter object would
693 use. See xshard.backend().
694 INT xshard_param.get_alt()
696 Get the alt parameter which a shard director using this parameter
697 object would use. See xshard.backend().
698 REAL xshard_param.get_warmup()
700 Get the warmup parameter which a shard director using this parameter
701 object would use. See xshard.backend().
702 BOOL xshard_param.get_rampup()
704 Get the rampup parameter which a shard director using this parameter
705 object would use. See xshard.backend().
706 STRING xshard_param.get_healthy()
708 Get a string representation of the healthy enum argument which a shard
709 director using this parameter object would use. See xshard.backend().
710 BLOB xshard_param.use()
712 This method may only be used in backend context.
713 For use with the param argument of xshard.backend() to associate this
715 shard parameter set with a shard director.
716 BACKEND lookup(STRING)
718 Lookup a backend by its name.
719 This function can only be used from vcl_init{} and vcl_fini{}.
721
723 Development of a previous version of the shard director was partly
724 sponsored by Deutsche Telekom AG - Products & Innovation.
725 Development of a previous version of the shard director was partly
727 sponsored by BILD GmbH & Co KG.
728
730 This document is licensed under the same licence as Varnish
731 itself. See LICENCE for details.
732 Copyright (c) 2013-2015 Varnish Software AS
734 Copyright 2009-2018 UPLEX - Nils Goroll Systemoptimierung
735 All rights reserved.
736 Authors: Poul-Henning Kamp <phk@FreeBSD.org>
738 Julian Wiesener <jw@uplex.de>
739 Nils Goroll <slink@uplex.de>
740 Geoffrey Simmons <geoff@uplex.de>
741 Redistribution and use in source and binary forms, with or without
743 modification, are permitted provided that the following conditions
744 are met:
745 1. Redistributions of source code must retain the above copyright
746 notice, this list of conditions and the following disclaimer.
747 2. Redistributions in binary form must reproduce the above copyright
748 notice, this list of conditions and the following disclaimer in the
749 documentation and/or other materials provided with the distribution.
750 THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
752 ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
753 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
754 ARE DISCLAIMED. IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
755 FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
756 DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
757 OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
758 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
759 LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
760 OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
761 SUCH DAMAGE.
762 3 VMOD(DIRECTORS)