Return codes

Each test within a testsuite returns one of the following return codes:

•

0 PASSED

•

1 FAILED

•

2 NOT TESTED - a test setup step or precondition check failed

•

3 SKIPPED - unmet requirements for testing

Note that a NOT TESTED result is treated as a failure of the entire test run, but SKIPPED is not.

The spectest and logintest shall return the same results whether they are run against a Mac AFP server or Netatalk.

Conformance tests

afp_spectest makes up the core of the AFP specification test suite, with several hundred test cases. It is organized into testsets, divided by the AFP commands tested, or by preconditions for testing. For instance, the tier 2 (T2) tests need to be run on the host with the -c option indicating the path to the shared volume. There are also read-only and sleep tests that need to be run separately.

afp_logintest is a testsuite for DSI sessions and authentication.

Benchmarking

afp_lantest is a file transfer benchmarking tool for AFP servers, inspired by HELIOS LanTest, which runs a batch of varied file transfer patterns.

afp_speedtest is a benchmark testsuite for read, write and copy operations. It can be run using either AFP commands or POSIX syscalls, handy for comparing netatalk speeds against other file transfer protocols.

Helpers

afparg is an interactive AFP client that takes an AFP command with optional arguments. This can be used for troubleshooting or system administration. Run afparg -l to list available commands.

fce_listen is a simple listener for Netatalk’s Filesystem Change Event (FCE) protocol. It will print out any UDP datagrams received from the AFP server.

Testing a Mac AFP server

This suite of tools were designed primarily to test Netatalk AFP servers, however they can also be used to test a native Mac OS AFP server hosted by an older Mac OS X or Classic Mac OS system.

Launch the test runner with the -m option when testing a Mac AFP server. When running in Mac mode, the test runner will report tests with known current or historical differences between Mac and Netatalk.

If Mac and Netatalk differ, or if Mac results differ between versions:

header.dsi_code       -5000     AFPERR_ACCESS
MAC RESULT: -5019 AFPERR_PARAM    -5010 AFPERR_BUSY
Netatalk returns AFPERR_ACCESS when a Mac return AFPERR_PARAM or AFPERR_BUSY

When Mac and Netatalk historically returned different results but now behave the same way:

Warning MAC and Netatalk now same RESULT!

Configure environment

Below is a sample configuration for running the AFP spec tests.

•

2 users: user1, user2 with the same password

•

1 group: afpusers

•

user1, user2 assigned to afpusers group

•

clear text UAM + guest UAM

Arrange two empty directories. Some tests will fail if there are residual files in the test directories.

drwxrwsr-x    5 user1   afpusers       176 avr 27 23:56 /tmp/afptest1
drwxrwsr-x    5 user1   afpusers       176 avr 27 23:56 /tmp/afptest2

Set afp.conf as follows:

[Global]
uam list = uams_clrtxt.so uams_guest.so
[testvol1]
ea = sys
path = /tmp/afptest1
valid users = @afpusers
volume name = testvol1
[testvol2]
ea = sys
path = /tmp/afptest2
valid users = @afpusers
volume name = testvol2

Running tests

Run the afp_spectest for the “FPSetForkParms_test” testset with AFP 3.4.

% afp_spectest -h 127.0.0.1 -p 548 -u user1 -d user2 -w passwd -s testvol1 -S testvol2 -c /srv/afptest1 -7 -f FPSetForkParms_test
===================
FPSetForkParms_test
-------------------
FPSetForkParms:test62: SetForkParams errors - PASSED
FPSetForkParms:test141: Setforkmode error - PASSED
FPSetForkParms:test217: Setfork size 64 bits - PASSED
FPSetForkParms:test306: set fork size, new size > old size - PASSED

Run the afp_lantest benchmark using AFP 3.0.

% afp_lantest -h 192.168.0.2 -s testvol1 -u user1 -w passwd -3
Run 0 => Opening, stating and reading 512 bytes from 1000 files   1799 ms
Run 0 => Writing one large file                                     30 ms for 100 MB (avg. 3495 MB/s)
Run 0 => Reading one large file                                      8 ms for 100 MB (avg. 13107 MB/s)
Run 0 => Locking/Unlocking 10000 times each                       1959 ms
Run 0 => Creating dir with 2000 files                             1339 ms
Run 0 => Enumerate dir with 2000 files                             217 ms
Run 0 => Create directory tree with 10^3 dirs                      496 ms
Netatalk Lantest Results (averages)
===================================
Opening, stating and reading 512 bytes from 1000 files   1799 ms
Writing one large file                                     30 ms for 100 MB (avg. 3495 MB/s)
Reading one large file                                      8 ms for 100 MB (avg. 13107 MB/s)
Locking/Unlocking 10000 times each                       1959 ms
Creating dir with 2000 files                             1339 ms
Enumerate dir with 2000 files                             217 ms
Create directory tree with 10^3 dirs                      496 ms