This document is for a development version of Ceph.
rbd-replay-many – replay a rados block device (RBD) workload on several clients¶
rbd-replay-many is a utility for replaying a rados block device (RBD) workload on several clients. Although all clients use the same workload, they replay against separate images. This matches normal use of librbd, where each original client is a VM with its own image.
Configuration and replay files are not automatically copied to clients. Replay images must already exist.
Specifies the name (and snap) of the originally traced image. Necessary for correct name mapping.
Prefix of image names to replay against. Specifying –image-prefix=foo results in clients replaying against foo-0, foo-1, etc. Defaults to the original image name.
Path to the rbd-replay executable.
Delay between starting each client. Defaults to 0.
rbd-replay-many host-0 host-1 --original-image=image -- -c ceph.conf replay.bin
This results in the following commands being executed:
ssh host-0 'rbd-replay' --map-image 'image=image-0' -c ceph.conf replay.bin ssh host-1 'rbd-replay' --map-image 'image=image-1' -c ceph.conf replay.bin
rbd-replay-many is part of Ceph, a massively scalable, open-source, distributed storage system. Please refer to the Ceph documentation at https://docs.ceph.com for more information.