2 # Copyright 1999-2007 The OpenLDAP Foundation, All Rights Reserved.
3 # COPYING RESTRICTIONS APPLY, see COPYRIGHT.
7 This is perhaps one of the most important chapters in the guide, because if
8 you have not tuned {{slapd}}(8) correctly or grasped how to design your
9 directory and environment, you can expect very poor performance.
11 Reading, understanding and experimenting using the instructions and information
12 in the following sections, will enable you to fully understand how to tailor
13 your directory server to your specific requirements.
15 It should be noted that the following information has been collected over time
16 from our community based FAQ. So obviously the benefit of this real world experience
17 and advice should be of great value to the reader.
20 H2: Performance Factors
22 Various factors can play a part in how your directory performs on your chosen
23 hardware and environment. We will attempt to discuss these here.
28 Scale your cache to use available memory and increase system memory if you can.
35 Use fast subsystems. Put each database and logs on separate disks.
37 Example showing config settings
42 http://www.openldap.org/faq/data/cache/363.html
47 H3: Directory Layout Design
49 Reference to other sections and good/bad drawing here.
59 http://www.openldap.org/faq/data/cache/42.html
60 http://www.connexitor.com/blog/pivot/entry.php?id=103#body
61 http://groups.google.com/group/comp.mail.sendmail/browse_frm/thread/17c5c0b94ad1fc58/f870758659375718?lnk=gst&q=hyc&rnum=12&hl=en#f870758659375718
66 http://www.openldap.org/faq/data/cache/80.html
69 H2: BDB/HDB Database Caching
71 We all know what caching is, don't we?
73 In brief, "A cache is a block of memory for temporary storage of data likely
74 to be used again" - {{URL:http://en.wikipedia.org/wiki/Cache}}
76 There are 3 types of caches, BerkeleyDB's own cache, {{slapd}}(8)
77 entry cache and {{TERM:IDL}} (IDL) cache.
82 BerkeleyDB's own data cache operates on page-sized blocks of raw data.
84 Note that while the {{TERM:BDB}} cache is just raw chunks of memory and
85 configured as a memory size, the {{slapd}}(8) entry cache holds parsed entries,
86 and the size of each entry is variable.
88 There is also an IDL cache which is used for Index Data Lookups.
89 If you can fit all of your database into slapd's entry cache, and all of your
90 index lookups fit in the IDL cache, that will provide the maximum throughput.
92 If not, but you can fit the entire database into the BDB cache, then you
93 should do that and shrink the slapd entry cache as appropriate.
95 Failing that, you should balance the BDB cache against the entry cache.
97 It is worth noting that it is not absolutely necessary to configure a BerkeleyDB
98 cache equal in size to your entire database. All that you need is a cache
99 that's large enough for your "working set."
101 That means, large enough to hold all of the most frequently accessed data,
102 plus a few less-frequently accessed items.
106 H4: Calculating Cachesize
108 The back-bdb database lives in two main files, {{F:dn2id.bdb}} and {{F:id2entry.bdb}}.
109 These are B-tree databases. We have never documented the back-bdb internal
110 layout before, because it didn't seem like something anyone should have to worry
111 about, nor was it necessarily cast in stone. But here's how it works today,
114 A B-tree is a balanced tree; it stores data in its leaf nodes and bookkeeping
115 data in its interior nodes (If you don't know what tree data structures look
116 like in general, Google for some references, because that's getting far too
117 elementary for the purposes of this discussion).
119 For decent performance, you need enough cache memory to contain all the nodes
120 along the path from the root of the tree down to the particular data item
121 you're accessing. That's enough cache for a single search. For the general case,
122 you want enough cache to contain all the internal nodes in the database.
126 will tell you how many internal pages are present in a database. You should
127 check this number for both dn2id and id2entry.
129 Also note that id2entry always uses 16KB per "page", while dn2id uses whatever
130 the underlying filesystem uses, typically 4 or 8KB. To avoid thrashing the,
131 your cache must be at least as large as the number of internal pages in both
132 the dn2id and id2entry databases, plus some extra space to accomodate the actual
135 For example, in my OpenLDAP 2.4 test database, I have an input LDIF file that's
136 about 360MB. With the back-hdb backend this creates a dn2id.bdb that's 68MB,
137 and an id2entry that's 800MB. db_stat tells me that dn2id uses 4KB pages, has
138 433 internal pages, and 6378 leaf pages. The id2entry uses 16KB pages, has 52
139 internal pages, and 45912 leaf pages. In order to efficiently retrieve any
140 single entry in this database, the cache should be at least
142 > (433+1) * 4KB + (52+1) * 16KB in size: 1736KB + 848KB =~ 2.5MB.
144 This doesn't take into account other library overhead, so this is even lower
145 than the barest minimum. The default cache size, when nothing is configured,
148 This 2.5MB number also doesn't take indexing into account. Each indexed attribute
149 uses another database file of its own, using a Hash structure.
151 Unlike the B-trees, where you only need to touch one data page to find an entry
152 of interest, doing an index lookup generally touches multiple keys, and the
153 point of a hash structure is that the keys are evenly distributed across the
154 data space. That means there's no convenient compact subset of the database that
155 you can keep in the cache to insure quick operation, you can pretty much expect
156 references to be scattered across the whole thing. My strategy here would be to
157 provide enough cache for at least 50% of all of the hash data.
159 > (Number of hash buckets + number of overflow pages + number of duplicate pages) * page size / 2.
161 The objectClass index for my example database is 5.9MB and uses 3 hash buckets
162 and 656 duplicate pages. So:
164 > ( 3 + 656 ) * 4KB / 2 =~ 1.3MB.
166 With only this index enabled, I'd figure at least a 4MB cache for this backend.
167 (Of course you're using a single cache shared among all of the database files,
168 so the cache pages will most likely get used for something other than what you
169 accounted for, but this gives you a fighting chance.)
171 With this 4MB cache I can slapcat this entire database on my 1.3GHz PIII in
172 1 minute, 40 seconds. With the cache doubled to 8MB, it still takes the same 1:40s.
173 Once you've got enough cache to fit the B-tree internal pages, increasing it
174 further won't have any effect until the cache really is large enough to hold
175 100% of the data pages. I don't have enough free RAM to hold all the 800MB
176 id2entry data, so 4MB is good enough.
178 With back-bdb and back-hdb you can use "db_stat -m" to check how well the
179 database cache is performing.
182 H3: {{slapd}}(8) Entry Cache
184 The {{slapd}}(8) entry cache operates on decoded entries. The rationale - entries
185 in the entry cache can be used directly, giving the fastest response. If an entry
186 isn't in the entry cache but can be extracted from the BDB page cache, that will
187 avoid an I/O but it will still require parsing, so this will be slower.
189 If the entry is in neither cache then BDB will have to flush some of its current
190 cached pages and bring in the needed pages, resulting in a couple of expensive
191 I/Os as well as parsing.
193 As far as balancing the entry cache vs the BDB cache - parsed entries in memory
194 are generally about twice as large as they are on disk.
196 As we have already mentioned, not having a proper database cache size will
197 cause performance issues. These issues are not an indication of corruption
198 occurring in the database. It is merely the fact that the cache is thrashing
199 itself that causes performance/response time to slowdown.
205 If you want to setup the cache size, please read:
207 (Xref) How do I configure the BDB backend?
208 (Xref) What are the DB_CONFIG configuration directives?
209 http://www.sleepycat.com/docs/utility/db_recover.html
211 A default config can be found in the answer:
213 (Xref) What are the DB_CONFIG configuration directives?
215 just change the set_lg_dir to point to your .log directory or comment that line.
218 - Create a DB_CONFIG file in your ldap home directory (/var/lib/ldap/DB_CONFIG) with the correct "set_cachesize" value
219 - stop your ldap server and run db_recover -h /var/lib/ldap
220 - start your ldap server and check the new cache size with:
222 db_stat -h /var/lib/ldap -m | head -n 2
224 - this procedure is only needed if you use OpenLDAP 2.2 with the BDB or HDB backends; In OpenLDAP 2.3 DB recovery is performed automatically whenever the DB_CONFIG file is changed or when an unclean shutdown is detected.
227 --On Tuesday, February 22, 2005 12:15 PM -0500 Dusty Doris <openldap@mail.doris.cc> wrote:
229 Few questions, if you change the cachesize and idlecachesize entries, do
230 you have to do anything special aside from restarting slapd, such as run
231 slapindex or db_recover?
234 Also, is there any way to tell how much memory these caches are taking up
235 to make sure they are not set too large? What happens if you set your
236 cachesize too large and you don't have enough available memory to store
237 these? Will that cause an issue with openldap, or will it just not cache
238 those entries that would make it exceed its available memory. Will it
239 just use some sort of FIFO on those caches?
242 It will consume the memory resources of your system, and likely cause issues.
244 Finally, what do most people try to achieve with these values? Would the
245 goal be to make these as big as the directory? So, if I have 400,000 dn's
246 in my directory, would it be safe to set these at 400000 or would
247 something like 20,000 be good enough to get a nice performance increase?
250 I try to cache the most actively used entries. Unless you expect all 400,000 entries of your DB to be accessed regularly, there is no need to cache that many entries. My entry cache is set to 20,000 (out of a little over 400,000 entries).
252 The idl cache has to do with how many unique result sets of searches you want to store in memory. Setting up this cache will allow your most frequently placed searches to get results much faster, but I doubt you want to try and cache the results of every search that hits your system. ;)
257 H3: {{TERM:IDL}} Cache
260 http://www.openldap.org/faq/data/cache/1076.html