PDF(100.5 KB) View with Adobe Reader on a variety of devices
Updated:July 17, 2015
The documentation set for this product strives to use bias-free language. For the purposes of this documentation set, bias-free is defined as language that does not imply discrimination based on age, disability, gender, racial identity, ethnic identity, sexual orientation, socioeconomic status, and intersectionality. Exceptions may be present in the documentation due to language that is hardcoded in the user interfaces of the product software, language used based on RFP documentation, or language that is used by a referenced third-party product. Learn more about how Cisco is using Inclusive Language.
If the 5012 diameter result code is observed at a high rate at higher TPS, proceed with the additional log verifications in this procedure.
Verify in the consolidated-engine log that the "connection wait timeout after 0 ms" error is observed before the Policy and Charging Rules Function (PCRF) sends the DiameterResponseMessage with Result-Code: 5012.
<snip> INFO : (balance) Error found, rolling back transaction ERROR : (core) Error processing policy request: com.mongodb.DBPortPool$Connection WaitTimeOut: Connection wait timeout after 0 ms com.mongodb.DBPortPool.get(DBPortPool.java:222) com.mongodb.DBTCPConnector$MyPort.get(DBTCPConnector.java:413) com.mongodb.DBTCPConnector.innerCall(DBTCPConnector.java:238) com.mongodb.DBTCPConnector.call(DBTCPConnector.java:216) com.mongodb.DBApiLayer$MyCollection.__find(DBApiLayer.java:288) com.mongodb.DBApiLayer$MyCollection.__find(DBApiLayer.java:273) com.mongodb.DBCollection.findOne(DBCollection.java:728) com.mongodb.DBCollection.findOne(DBCollection.java:708) com.broadhop.balance.impl.dao.impl.MongoBalanceRepository$6.findOne(MongoBalance Repository.java:375) <snip>
Note: You can check TPS on the CPS system in the middle of a problematic time with the top_qps.sh command that is available in CPS version 5.5 and later.
Change the Threading configuration in Policy Builder from default 20 to 50. In order to do this, login to Policy Builder and choose Reference Data > Systems > system-1 > Plugin Configurations >Threading Configurations.
By default (when the Threading Configuration field is blank) the number of threads for mongo connection is 20 in Policy Builder configuration so it can handle that amount of requests when it runs on low TPS. As the TPS increases these threads are busy and hence more threads are required in order to fulfill the requests.
Thread count of 50 is sufficient in order to handle around 5000 TPS as more threads are available that can handle a higher number of requests.
These are policy engine threads and are defined with the name "rules"and should be configured with that name only.
Add Dmongo.client.thread.maxWaitTime=5000 to /etc/broadhop/pcrf/qns.conf.
Dmongo.client.thread.maxWaitTime is a time in milliseconds a thread waits for a connection to become available. If this parameter is not specified it considers the default value which is 0 ms. Therefore, the error is observed while tests are at a higher TPS. The addition of this parameter in /etc/broadhop/pcrf/qns.conf increases the time the new threads wait for mongo connection when tests are on a high TPS.
2000 is the QA recommended value and was tested for high TPS. For TPS greater than 5000 you can configure it to 5000 ms in order to optimize the performance.
Add -Dspr.mongo.socket.timeout=5000 to /etc/broadhop/pcrf/qns.conf.
By default the value is 60000 milliseconds.(60 seconds). It therefore takes longer time to become available for other threads.
The recommended configuration is 5000 milliseconds (5 seconds) in order to facilitate a quicker timeout and allow other threads to process fast.
Change the Connections Per Host value from default 5 to 20 in Policy Builder. In order ot do this, login to Policy Builder and choose Reference Data > Systems > system-1 > Plugin Configurations > USuM Configuration > Connections Per Host.
This is the per Quantum Network Suite (QNS) number of connections with mongo DB. This means that for 4 QNS, 4*20=80 is the total number of connections.
This is required for frequent updates in mongodb. Therefore, it is recommended to be updated as 20 per QA recommendation for optimal performance.
Also configure Db Read Preference as SecondaryPreferred which means that all the QNS receives data from the Secondary database and only receives data from Primary when the Secondary DB is busy. This helps to optimize the performance since Primary DB is least loaded.
Configure the appropriate root logging level for the System.
Excessive logs can block processing on the QNS and LB level.Therefore it is recommended that you configure the root logging level at warn or higher levels both at /etc/brodhop/logback.xml and /etc/broadhop/controlcenter/logback.xml files.