Mercurial > hg > nginx-site
comparison xml/en/docs/http/ngx_http_upstream_module.xml @ 1126:2e28d1c4a0a2
Upstream: revised layout of the "server" directive.
author | Ruslan Ermilov <ru@nginx.com> |
---|---|
date | Fri, 21 Mar 2014 10:55:50 +0400 |
parents | 9a95b9026a16 |
children | 948c853f5d22 |
comparison
equal
deleted
inserted
replaced
1125:9a95b9026a16 | 1126:2e28d1c4a0a2 |
---|---|
8 <!DOCTYPE module SYSTEM "../../../../dtd/module.dtd"> | 8 <!DOCTYPE module SYSTEM "../../../../dtd/module.dtd"> |
9 | 9 |
10 <module name="Module ngx_http_upstream_module" | 10 <module name="Module ngx_http_upstream_module" |
11 link="/en/docs/http/ngx_http_upstream_module.html" | 11 link="/en/docs/http/ngx_http_upstream_module.html" |
12 lang="en" | 12 lang="en" |
13 rev="17"> | 13 rev="18"> |
14 | 14 |
15 <section id="summary"> | 15 <section id="summary"> |
16 | 16 |
17 <para> | 17 <para> |
18 The <literal>ngx_http_upstream_module</literal> module | 18 The <literal>ngx_http_upstream_module</literal> module |
98 <example> | 98 <example> |
99 upstream backend { | 99 upstream backend { |
100 server backend1.example.com weight=5; | 100 server backend1.example.com weight=5; |
101 server 127.0.0.1:8080 max_fails=3 fail_timeout=30s; | 101 server 127.0.0.1:8080 max_fails=3 fail_timeout=30s; |
102 server unix:/tmp/backend3; | 102 server unix:/tmp/backend3; |
103 | |
104 server backup1.example.com backup; | |
103 } | 105 } |
104 </example> | 106 </example> |
105 </para> | 107 </para> |
106 | 108 |
107 <para> | 109 <para> |
176 | 178 |
177 </list> | 179 </list> |
178 By default, the parameter is set to 10 seconds. | 180 By default, the parameter is set to 10 seconds. |
179 </tag-desc> | 181 </tag-desc> |
180 | 182 |
183 <tag-name><literal>backup</literal></tag-name> | |
184 <tag-desc> | |
185 marks the server as a backup server. | |
186 It will be passed requests when the primary servers are unavailable. | |
187 </tag-desc> | |
188 | |
189 <tag-name><literal>down</literal></tag-name> | |
190 <tag-desc> | |
191 marks the server as permanently unavailable; used along with | |
192 the <link id="ip_hash"/> directive. | |
193 </tag-desc> | |
194 | |
195 </list> | |
196 </para> | |
197 | |
198 <para> | |
199 Additionally, | |
200 the following parameters are available as part of our <commercial_version/>: | |
201 <list type="tag"> | |
202 | |
203 <tag-name id="max_conns"> | |
204 <literal>max_conns</literal>=<value>number</value> | |
205 </tag-name> | |
206 <tag-desc> | |
207 limits the maximum <value>number</value> of simultaneous connections to the | |
208 proxied server (1.5.9). | |
209 Default value is zero, meaning there is no limit. | |
210 </tag-desc> | |
211 | |
212 <tag-name id="resolve"> | |
213 <literal>resolve</literal> | |
214 </tag-name> | |
215 <tag-desc> | |
216 monitors changes of the IP addresses | |
217 that correspond to a domain name of the server, | |
218 and automatically modifies the upstream configuration | |
219 without the need of restarting nginx (1.5.12) | |
220 <para> | |
221 In order for this parameter to work, | |
222 the <link doc="ngx_http_core_module.xml" id="resolver"/> directive | |
223 must be specified in the | |
224 <link doc="ngx_http_core_module.xml" id="http"/> block. | |
225 <example> | |
226 http { | |
227 resolver 10.0.0.1; | |
228 | |
229 upstream u { | |
230 zone ...; | |
231 ... | |
232 server example.com resolve; | |
233 } | |
234 } | |
235 </example> | |
236 </para> | |
237 </tag-desc> | |
238 | |
239 <tag-name><literal>route</literal>=<value>string</value></tag-name> | |
240 <tag-desc> | |
241 sets the server route name. | |
242 </tag-desc> | |
243 | |
181 <tag-name><literal>slow_start</literal>=<value>time</value></tag-name> | 244 <tag-name><literal>slow_start</literal>=<value>time</value></tag-name> |
182 <tag-desc> | 245 <tag-desc> |
183 sets the <value>time</value> during which the server will recover its weight | 246 sets the <value>time</value> during which the server will recover its weight |
184 from zero to a nominal value, when unhealthy server becomes | 247 from zero to a nominal value, when unhealthy server becomes |
185 <link id="health_check">healthy</link>, | 248 <link id="health_check">healthy</link>, |
186 or when the server becomes available after a period of time | 249 or when the server becomes available after a period of time |
187 it was considered unavailable. | 250 it was considered unavailable. |
188 Default value is zero, i.e. slow start is disabled. | 251 Default value is zero, i.e. slow start is disabled. |
189 <note> | |
190 This functionality is available as part of our <commercial_version/>. | |
191 </note> | |
192 </tag-desc> | |
193 | |
194 <tag-name id="max_conns"> | |
195 <literal>max_conns</literal>=<value>number</value> | |
196 </tag-name> | |
197 <tag-desc> | |
198 limits the maximum <value>number</value> of simultaneous connections to the | |
199 proxied server (1.5.9). | |
200 Default value is zero, meaning there is no limit. | |
201 <note> | |
202 This functionality is available as part of our <commercial_version/> only. | |
203 </note> | |
204 </tag-desc> | |
205 | |
206 <tag-name><literal>backup</literal></tag-name> | |
207 <tag-desc> | |
208 marks the server as a backup server. | |
209 It will be passed requests when the primary servers are unavailable. | |
210 </tag-desc> | |
211 | |
212 <tag-name><literal>down</literal></tag-name> | |
213 <tag-desc> | |
214 marks the server as permanently unavailable; used along with | |
215 the <link id="ip_hash"/> directive. | |
216 </tag-desc> | |
217 | |
218 <tag-name id="resolve"> | |
219 <literal>resolve</literal> | |
220 </tag-name> | |
221 <tag-desc> | |
222 monitors changes of the IP addresses | |
223 that correspond to a domain name of the server, | |
224 and automatically modifies the upstream configuration | |
225 without the need of restarting nginx (1.5.12) | |
226 <para> | |
227 In order for this parameter to work, | |
228 the <link doc="ngx_http_core_module.xml" id="resolver"/> directive | |
229 must be specified in the | |
230 <link doc="ngx_http_core_module.xml" id="http"/> block. | |
231 <example> | |
232 http { | |
233 resolver 10.0.0.1; | |
234 | |
235 upstream u { | |
236 zone ...; | |
237 ... | |
238 server example.com resolve; | |
239 } | |
240 } | |
241 </example> | |
242 </para> | |
243 <note> | |
244 This functionality is available as part of our <commercial_version/>. | |
245 </note> | |
246 </tag-desc> | |
247 | |
248 <tag-name><literal>route</literal>=<value>string</value></tag-name> | |
249 <tag-desc> | |
250 sets the server route name. | |
251 <note> | |
252 This functionality is available as part of our <commercial_version/>. | |
253 </note> | |
254 </tag-desc> | 252 </tag-desc> |
255 | 253 |
256 </list> | 254 </list> |
257 </para> | 255 </para> |
258 | 256 |
259 <para> | 257 <para> |
260 Example: | |
261 <example> | |
262 upstream backend { | |
263 server backend1.example.com weight=5; | |
264 server 127.0.0.1:8080 max_fails=3 fail_timeout=30s; | |
265 server unix:/tmp/backend3; | |
266 | |
267 server backup1.example.com:8080 backup; | |
268 } | |
269 </example> | |
270 | |
271 <note> | 258 <note> |
272 If there is only a single server in a group, <literal>max_fails</literal>, | 259 If there is only a single server in a group, <literal>max_fails</literal>, |
273 <literal>fail_timeout</literal> and <literal>slow_start</literal> parameters | 260 <literal>fail_timeout</literal> and <literal>slow_start</literal> parameters |
274 are ignored, and such a server will never be considered unavailable. | 261 are ignored, and such a server will never be considered unavailable. |
275 </note> | 262 </note> |