SOLR distributed deployment

The SOLR replication mode is an implementation method used to synchronize master-slave servers in a distributed environment. Because the deployment cost of the rsync-based SOLR is too high, replaced by solr1.4, it is replaced by an HTTP-based index file transfer mechanism. This method is easy to deploy and requires only one file.

The procedure is as follows:

The steps are divided into the master server and slave server. Multiple slave servers are allowed, that is, the configuration of the slave server is the same.

Master Server:

 

In solrconfig. XML, find the following line:

<Requesthandler name = "/replication" class = "SOLR. replicationhandler">

1. <Lst name = "master">
2. <STR name = "replicateafter"> commit </STR>
4. <STR name = "conffiles"> schema.xml,stopwords.txt, elevate. xml </STR>
6. <STR name = "commitreserveduration"> 00:05:00 </STR>
7. <STR name = "httpbasicauthuser"> 123 </STR>
8. <STR name = "httpbasicauthpassword"> 123 </STR>
11. </Lst>
12. </Requesthandler>

 

Note:

• Replicateafter: SOLR will perform a copy after the following operations occur: 'commit ', 'startup' 'optimize'. Here we choose commit, that is, after SOLR receives a commit request every time, the replication policy is executed.
• Conffiles: the configuration file to be distributed. SOLR also synchronizes the field configuration files schema.xmland stopwords.txt on the master server, and the solid partition file: elevate. XML to the slave server.
• Commitreserveduration: the cycle time for retaining the incremental index after each commit. Set this parameter to 5 minutes.

Slave Server

<Requesthandler name = "/replication" class = "SOLR. replicationhandler">

1. <Lst name = "slave">
3. <STR name = "masterurl"> http: // localhost: Port/SOLR/corename/replication </STR>
5. <STR name = "pollinterval"> 00:05:00 </STR>
7. <STR name = "compression"> internal </STR>
9. <STR name = "httpconntimeout"> 5000 </STR>
10. <STR name = "httpreadtimeout"> 10000 </STR>
13. <STR name = "httpbasicauthuser"> 123 </STR>
14. <STR name = "httpbasicauthpassword"> 123 </STR>
16. </Lst>
17. </Requesthandler>

 

Note:

• Masterurl: synchronization URL of the master server
• Pollinterval: The slave server synchronization interval, that is, the interval at which the master server Synchronizes time.
• Httpconntimeout: sets the connection timeout (unit: milliseconds)
• Httpreadtimeout: If the synchronization index file is too large, increase this value as appropriate. (Unit: milliseconds)
• Httpbasicauthuser: the authentication username, which must be consistent with the master server
• Httpbasicauthpassword: The verification password, which must be consistent with the master server
• Compression: external or internal uses SOLR's own compression algorithm or application container's

Difference: the internal algorithm will greatly increase the synchronization cost. original statement: use this only if your bandwidth is low. This can actually slowdown replication in a LAN.

Therefore, we recommend that you use the external method: external.

At the same time, you need to configure the external application container: Take Tomcat as an example:

1. <Connector compression = "on"
2. Compressablemimetype = "text/html, text/XML, text/plain"
3. Compressionminsize = "2048" type = "parmname" text = "parmname"/>