NAME

SYNOPSIS

OPTIONS

--cleanup!

Defaults to true. Negate with --nocleanup. If true, all generated files are removed at the end of the test run.

--files=i

Number of files to run through the experiment. Default is 15.

--help|h!

This help

--sleep1=f

Defaults to 0.2. Seconds to sleep between the cration of the initial files.

--sleep2=f

Defaults to 0.1. Seconds to sleep between the iterations of the second phase.

--iterations=i

Defaults to 30. Number of iterations in the second phase.

DESCRIPTION

In the first phase the test creates a couple of files and injects them into the tree, one after the other. There are tunable sleep1 pauses between each file creation. In the second phase the test runs alternating aggregate commands on the server and rmirror commands on the client. After each iteration both directories are checksummed and stored in a separate yaml file for later inspection.

If you want to inspect the yaml files, be sure to set --nocleanup.

Interpretation of the output

Output may look like this:

# 17.1575 new state reached in t/serv-5c59696a590715c20f2b7f55c281c667.yaml
# 18.0686 new state reached in t/mirr-b9b903e62f31249d2d5836eede1d0420.yaml
# 19.2339 new state reached in t/serv-9a9df7f3c8d2fc501c27490696ba1c88.yaml
# 33.2662 new state reached in t/serv-7ad22e96a3ecf527e1fa934425ec7516.yaml
# 55.2330 new state reached in t/serv-ce628a7ee14eb32054f6744ab9772b2c.yaml

This means that the RECENT files on the server have changed 4 times due to calls to aggregate but the RECENT files on the mirror have only changed once.