Scripts for automation of I/O workload and Ceph failure injection. Uses COSbench to apply I/O workload; pbench for monitoring and SSH scripts (dropOSD.bash) to inject failures.
For a writeup of how the ceph cluster was installed see: https://github.com/ekaynar/Benchmarks/blob/master/ceph-ansible/README.md
FILE INVENTORY:
- vars.shinc - global variables
- writeXMLs.sh
- prepCluster.sh
- runtest.sh - main driver script which executes COSbench and injects failures
- XMLtemplates (directory)
- TMPL_deletewrite.xml
- TMPL_prepCluster.xml
- TMPL_hybrid.xml
- Utils (directory)
- dropOSD.bash - script which is run on OSDhostname (drops an OSD device)
- functions.shinc - collection of functions
- pollceph.sh - script run on MONhostname (polls ceph status)
USAGE:
- Edit 'vars.shinc' for your environment (hostnames; runtime; obj sizes ...)
- run 'writeXMLs.sh' <-- create COSbench workload files
- run 'prepCluster.sh' <-- create pools and fill the cluster
- run 'runtest.sh' <-- run the test and record results