scripts to investigate RGW garbage collection rates
Uses COSbench to generate RGW operations. All three scripts report on %RAW USED and Pending GC's All scripts create logfiles in RESULTS directory
- writeXML.sh writes the three XML files from the Templates (found in 'XMLtemplates' dir)
- resetRGW.sh resets the RGW env. Deletes pools and creates new user. Inserts passwd into XML files
- emptyCluster.sh invokes emptyWorkload.xml (runs cleanup and dispose operations)
- fillCluster.sh invokes fillWorkload.xml
- runIOworkload.sh invokes ioWorkload.xml
- copyPasswd.sh inserts the RGW password into the three XML workload files
- Utils/pollGC.sh called by runIOworkload.sh to periodically report stats
NOTE: host IPaddresses and ceph login credentials in vars.shinc will need to be replaced for your cluster
- edit vars.shinc
- writeXML.sh <-- you must run either 'resetRGW.sh' or 'copyPasswd.sh'
- resetRGW.sh
- fillCluster.sh
- runIOworkload.sh
- emptyCluster.sh
Based on a cluster with these parameters:
- 174TB of raw capacity
- 10 buckets/containers
- 1MB objext size
- three way replication the following object count (per container) produces the following fill levels:
- 25% full == 1.45M obj per container, 14.5M total objects
- 50% full == 2.9M objects per container, 29M total objects
Those values were calculated using this formula:
- Raw Capacity of 174TB / 3 = 58TB (factor in 3way repl)
- 58TB / 10 = 5.8TB (factor in 10 containers)
- 5.8TB / 1MB = 5.8M object count capacity per container
- 25% of 5.8M is 1.45M objects per container
- 50% of 5.8M is 2.9M objects per container