Introduction
This document describes how to configure the threshold for fragmentation of the Balance database (DB).
Prerequisites
Requirements
Cisco recommends that you have knowledge of these topics:
- Linux
- Cisco Policy Suite (CPS)
- MongoDB
Components Used
The information in this document is based on these software and hardware versions:
The information in this document was created from the devices in a specific lab environment. All of the devices used in this document started with a cleared (default) configuration. If your network is live, ensure that you understand the potential impact of any command.
Background Information
Two application patterns cause increased Fragmentation, Document growth, and Document removal. The two operations are usual at the MongoDB level. The best way to defragment is to resync the data across replica members. However, there is still a fragmentation alarm.
- Fragmentation percent threshold values are configured in /etc/collectd.d/dbMonitorList.cfg file (present on sessionmgr Virtual Machines) for all the databases. The default threshold value for all the databases is configured as 40%. The default fragmentation threshold value can be changed as required. For more information, refer to Configure Custom Database Fragmentation Threshold Percentage section in the CPS Operations Guide.
- Additionally, verify the current fragmentation percentage of primary members of session_cache, sk_cache, diameter, and Subscriber Profile Repository (SPR) Database with the diagnostics.sh --get_frag_status command.
- The diagnostics.sh --get_frag_status supports fragmentation percentage calculation of session_cache, sk_cache, diameter, and the SPR Databases.
- The diagnostics.sh --get_frag_status like diagnostics.sh --get_session_shard_health is supported with root user only for multi-cluster High Availability (HA) and Geo-Redundant (GR) setups.
Steps
Pre-check
Note: All the commands are run from cluman.
Get the current Balance DB Fragmentation threshold, that is used in the verification stage.
# for host in $(hosts-all.sh | grep 'sessionmgr'); do echo check in progress on $host; ssh $host "cat /etc/collectd.d/dbMonitorList.cfg | grep balance"; done
Apply the Change
Add this entry in /var/qps/config/deploy/csv/Configuration.csv with the threshold value (change 50 to the required value, for example, 60 on normal fragmentation level).
balance_mgmt,50,
echo -e "\nbalance_mgmt,50," >> /var/qps/config/deploy/csv/Configuration.csv
[root@cps194cluman ]# echo -e "\nbalance_mgmt,50," >> /var/qps/config/deploy/csv/Configuration.csv
[root@cps194cluman ]# grep balance_mgmt Configuration.csv
balance_mgmt,50,
Import the csv Changes
/var/qps/install/current/scripts/import/import_deploy.sh
[root@cps194cluman csv]# /var/qps/install/current/scripts/import/import_deploy.sh
Filenames to be processed are listed here.
AdditionalHosts.csv Configuration.csv DBConfigServer.csv Definitions.csv Hosts.csv ReplicationSets.csv SessionCache.csv VLANs.csv VMSpecification.csv SecureConfig.csv VipProxyConfiguration.csv DSCPConfig.csv CriticalFiles.csv
Warning: The first build is not processed. We process only the list of files mentioned.
Warning: old files are not processed. We process only list of files mentioned.
The CSV files in /var/qps/config/deploy/csv are converted to json files in /var/qps/config/deploy/json..
build the hosts file to /var/www/html/hosts...
build the /etc/hosts file from the json configuation... /etc/hosts is backed to /etc/hosts.back
Rotate backup '/etc/hosts.back' to '/etc/hosts.back.0'
Backed up '/etc/hosts' to '/etc/hosts.back'
‘/etc/hosts’ -> ‘/etc/hosts.back’
Redis by default disabled -DenableQueueSystem=false in /etc/broadhop/qns.conf
Removing feature configs moved to core
Removing ws feature from pb and pcrf feature file
Building /etc/broadhop...
Copying to /var/qps/images/etc.tar.gz...
Creating MD5 Checksum...
Generating /etc/broadhop/servers.all
Rebuilding facts for: 'installer' (aka 'installer')
Creating md5sum for hosts file to validate later
Rebuilding facts for: 'casant01-ps01' (aka 'qns01')
Rebuilding facts for: 'casant01-pd02' (aka 'lb02')
Rebuilding facts for: 'casant01-sessionmgr01' (aka 'sessionmgr01')
Rebuilding facts for: 'casant01-sessionmgr02' (aka 'sessionmgr02')
Rebuilding facts for: 'casant01-oam01' (aka 'pcrfclient01')
Rebuilding facts for: 'casant01-ps02' (aka 'qns02')
Rebuilding facts for: 'casant01-pd01' (aka 'lb01')
Rebuilding facts for: 'casant01-oam02' (aka 'pcrfclient02')
Copying /etc/puppet to /var/qps/images/puppet.tar.gz...
Creating MD5 Checksum...
[root@cps194cluman csv]#
Build Updated Virtual Machine Images
# /var/qps/bin/build/build_all.sh
Update the Threshold in sessionmgr Virtual Machines
# /var/qps/install/current/scripts/upgrade/reinit.sh
Verification
Verify the threshold has been changed for the Balance DB fragmentation to the configured value.
# for host in $(hosts-all.sh | grep 'sessionmgr'); do echo check in progress on $host; ssh $host "cat /etc/collectd.d/dbMonitorList.cfg | grep balance"; done
Verify the alarm generation script contains the updated threshold.
# for host in $(hosts-all.sh | grep 'sessionmgr'); do echo checking in $host; ssh $host "cat /var/log/broadhop//scripts/gen-frag-trap.log | grep DEBUG | tail -5"; done
Note: Test this in a laboratory before you attempt production setup.
Feedback