%%
%% %CopyrightBegin%
%%
%% Copyright Ericsson AB 1996-2011. All Rights Reserved.
%%
%% The contents of this file are subject to the Erlang Public License,
%% Version 1.1, (the "License"); you may not use this file except in
%% compliance with the License. You should have received a copy of the
%% Erlang Public License along with this software. If not, it can be
%% retrieved online at http://www.erlang.org/.
%%
%% Software distributed under the License is distributed on an "AS IS"
%% basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
%% the License for the specific language governing rights and limitations
%% under the License.
%%
%% %CopyrightEnd%
%%
-module(disksup_SUITE).
-include_lib("test_server/include/test_server.hrl").
%% Test server specific exports
-export([all/0, suite/0,groups/0,init_per_group/2,end_per_group/2]).
-export([init_per_suite/1, end_per_suite/1]).
-export([init_per_testcase/2, end_per_testcase/2]).
%% Test cases
-export([api/1, config/1, alarm/1]).
-export([port/1]).
-export([terminate/1, unavailable/1, restart/1]).
-export([otp_5910/1]).
%% Default timetrap timeout (set in init_per_testcase)
-define(default_timeout, ?t:minutes(1)).
init_per_suite(Config) when is_list(Config) ->
ok = application:start(os_mon),
Config.
end_per_suite(Config) when is_list(Config) ->
ok = application:stop(os_mon),
Config.
init_per_testcase(unavailable, Config) ->
terminate(Config),
init_per_testcase(dummy, Config);
init_per_testcase(_Case, Config) ->
Dog = ?t:timetrap(?default_timeout),
[{watchdog,Dog} | Config].
end_per_testcase(unavailable, Config) ->
restart(Config),
end_per_testcase(dummy, Config);
end_per_testcase(_Case, Config) ->
Dog = ?config(watchdog, Config),
?t:timetrap_cancel(Dog),
ok.
suite() -> [{ct_hooks,[ts_install_cth]}].
all() ->
Bugs = [otp_5910],
case test_server:os_type() of
{unix, sunos} ->
[api, config, alarm, port, unavailable] ++ Bugs;
{unix, _OSname} -> [api, alarm] ++ Bugs;
{win32, _OSname} -> [api, alarm] ++ Bugs;
_OS -> [unavailable]
end.
groups() ->
[].
init_per_group(_GroupName, Config) ->
Config.
end_per_group(_GroupName, Config) ->
Config.
api(suite) -> [];
api(doc) -> ["Test of API functions"];
api(Config) when is_list(Config) ->
%% get_disk_data()
Data = disksup:get_disk_data(),
case lists:keysearch("/ldisk", 1, Data) of
{value,{Id,KByte,Capacity}} ->
true = io_lib:printable_list(Id),
true = is_integer(KByte),
true = is_integer(Capacity),
true = Capacity>0,
true = KByte>0;
_ ->
[{Id, KByte, Capacity}|_] = Data,
true = io_lib:printable_list(Id),
true = is_integer(KByte),
true = is_integer(Capacity),
% can be zero
% true = Capacity>0,
true = KByte>0
end,
%% get_check_interval()
1800000 = disksup:get_check_interval(),
%% set_check_interval(Minutes)
ok = disksup:set_check_interval(20),
1200000 = disksup:get_check_interval(),
{'EXIT',{badarg,_}} = (catch disksup:set_check_interval(0.5)),
1200000 = disksup:get_check_interval(),
ok = disksup:set_check_interval(30),
%% get_almost_full_threshold()
80 = disksup:get_almost_full_threshold(),
%% set_almost_full_threshold(Float)
ok = disksup:set_almost_full_threshold(0.90),
90 = disksup:get_almost_full_threshold(),
{'EXIT',{badarg,_}} =
(catch disksup:set_almost_full_threshold(-0.5)),
90 = disksup:get_almost_full_threshold(),
ok = disksup:set_almost_full_threshold(0.80),
ok.
config(suite) -> [];
config(doc) -> ["Test configuration"];
config(Config) when is_list(Config) ->
%% Change configuration parameters and make sure change is reflected
%% when disksup is restarted
ok = application:set_env(os_mon, disk_space_check_interval, 29),
ok = application:set_env(os_mon, disk_almost_full_threshold, 0.81),
ok = supervisor:terminate_child(os_mon_sup, disksup),
{ok, _Child1} = supervisor:restart_child(os_mon_sup, disksup),
1740000 = disksup:get_check_interval(),
81 = disksup:get_almost_full_threshold(),
%% Also try this with bad parameter values, should be ignored
ok =
application:set_env(os_mon, disk_space_check_interval, 0.5),
ok =
application:set_env(os_mon, disk_almost_full_threshold, -0.81),
ok = supervisor:terminate_child(os_mon_sup, disksup),
{ok, _Child2} = supervisor:restart_child(os_mon_sup, disksup),
1800000 = disksup:get_check_interval(),
80 = disksup:get_almost_full_threshold(),
%% Reset configuration parameters
ok = application:set_env(os_mon, disk_space_check_interval, 30),
ok = application:set_env(os_mon, disk_almost_full_threshold, 0.80),
ok.
%%----------------------------------------------------------------------
%% NOTE: The test case is a bit weak as it will fail if the disk usage
%% changes too much during its course, or if there are timing problems
%% with the alarm_handler receiving the alarms too late
%%----------------------------------------------------------------------
alarm(suite) -> [];
alarm(doc) -> ["Test that alarms are set and cleared"];
alarm(Config) when is_list(Config) ->
%% Find out how many disks exceed the threshold
%% and make sure the corresponding number of alarms is set
Threshold1 = disksup:get_almost_full_threshold(), % 80
Data1 = disksup:get_disk_data(),
Over1 = over_threshold(Data1, Threshold1),
Alarms1 = get_alarms(),
if
Over1==length(Alarms1) ->
true;
true ->
dump_info(),
?t:fail({bad_alarms, Threshold1, Data1, Alarms1})
end,
%% Try to find a disk with space usage below Threshold1,
%% lower the threshold accordingly and make sure new alarms are set
Fun1 = fun({_Id, _Kbyte, Capacity}) ->
if
Capacity>0, Capacity<Threshold1 -> true;
true -> false
end
end,
case until(Fun1, Data1) of
{_, _, Cap1} ->
Threshold2 = Cap1-1,
ok =
disksup:set_almost_full_threshold(Threshold2/100),
disksup ! timeout, % force a disk check
Data2 = disksup:get_disk_data(),
Over2 = over_threshold(Data2, Threshold2),
Alarms2 = get_alarms(),
if
Over2==length(Alarms2), Over2>Over1 ->
true;
true ->
dump_info(),
?t:fail({bad_alarms, Threshold2, Data2, Alarms2})
end;
false ->
ignore
end,
%% Find out the highest space usage among all disks
%% and try to raise the threshold above this value,
%% make sure all alarms are cleared
Fun2 = fun({_Id, _Kbyte, Capacity}, MaxAcc) ->
if
Capacity>MaxAcc -> Capacity;
true -> MaxAcc
end
end,
case lists:foldl(Fun2, 0, Data1) of
Max when Max<100 ->
Threshold3 = Max+1,
ok = disksup:set_almost_full_threshold(Threshold3/100),
disksup ! timeout, % force a disk check
Data3 = disksup:get_disk_data(),
Over3 = over_threshold(Data3, Threshold3),
Alarms3 = get_alarms(),
if
Over3==0, length(Alarms3)==0 ->
ok;
true ->
dump_info(),
?t:fail({bad_alarms, Threshold3, Data3, Alarms3})
end;
100 ->
ignore
end,
%% Reset threshold
ok = disksup:set_almost_full_threshold(Threshold1/100),
ok.
over_threshold(Data, Threshold) ->
Data2 = remove_duplicated_disks(lists:keysort(1, Data)),
lists:foldl(fun
({_Id, _Kbyte, Cap}, N) when Cap>=Threshold -> N+1;
(_DiskData, N) -> N
end, 0, Data2).
%% On some platforms (for example MontaVista) data for one disk can be
%% "duplicated":
%% Linux ppb 2.4.20_mvl31-pcore680 #1 Sun Feb 1 23:12:56 PST 2004 ppc unknown
%%
%% MontaVista(R) Linux(R) Professional Edition 3.1
%%
%% [ppb:~]> /bin/df -lk
%% Filesystem 1k-blocks Used Available Use% Mounted on
%% rootfs 8066141 3023763 4961717 38% /
%% /dev/root 8066141 3023763 4961717 38% /
%% tmpfs 192892 0 192892 0% /dev/shm
%%
%% disksup:
%% [{"/",8066141,38}, {"/",8066141,38}, {"/dev/shm",192892,0}]
%%
%% disksup will only set ONE alarm for "/".
%% Therefore the list of disk data must be sorted and duplicated disk
%% tuples removed before calculating how many alarms should be set, or
%% the testcase will fail erroneously.
remove_duplicated_disks([{Id, _, _}, {Id, Kbyte, Cap}|T]) ->
remove_duplicated_disks([{Id, Kbyte, Cap}|T]);
remove_duplicated_disks([H|T]) ->
[H|remove_duplicated_disks(T)];
remove_duplicated_disks([]) ->
[].
get_alarms() ->
lists:filter(fun
({{disk_almost_full, _Disk},_}) -> true;
(_) -> false
end, alarm_handler:get_alarms()).
until(Fun, [H|T]) ->
case Fun(H) of
true -> H;
false -> until(Fun, T)
end;
until(_Fun, []) -> false.
port(suite) -> [];
port(doc) ->
["Test that disksup handles a terminating port program"];
port(Config) when is_list(Config) ->
Str = os:cmd("ps -ef | grep '[d]isksup'"),
case io_lib:fread("~s ~s", Str) of
{ok, [_Uid,Pid], _Rest} ->
%% Monitor disksup
MonRef = erlang:monitor(process, disksup),
[{_Disk1,Kbyte1,_Cap1}|_] = disksup:get_disk_data(),
true = Kbyte1>0,
%% Kill the port program
case os:cmd("kill -9 " ++ Pid) of
[] ->
%% disksup should now terminate
receive
{'DOWN', MonRef, _, _, {port_died, _Reason}} ->
ok;
{'DOWN', MonRef, _, _, Reason} ->
?t:fail({unexpected_exit_reason, Reason})
after
3000 ->
?t:fail({still_alive, Str})
end,
%% Give os_mon_sup time to restart disksup
?t:sleep(?t:seconds(3)),
[{_Disk2,Kbyte2,_Cap2}|_] =
disksup:get_disk_data(),
true = Kbyte2>0,
ok;
Line ->
erlang:demonitor(MonRef),
{skip, {not_killed, Line}}
end;
_ ->
{skip, {os_pid_not_found, Str}}
end.
terminate(suite) -> [];
terminate(Config) when is_list(Config) ->
ok = application:set_env(os_mon, start_disksup, false),
ok = supervisor:terminate_child(os_mon_sup, disksup),
ok.
unavailable(suite) -> [];
unavailable(doc) ->
["Test correct behaviour when service is unavailable"];
unavailable(Config) when is_list(Config) ->
%% Make sure all API functions return their dummy values
[{"none",0,0}] = disksup:get_disk_data(),
1800000 = disksup:get_check_interval(),
ok = disksup:set_check_interval(5),
80 = disksup:get_almost_full_threshold(),
ok = disksup:set_almost_full_threshold(0.9),
ok.
restart(suite) ->
[];
restart(Config) when is_list(Config) ->
ok = application:set_env(os_mon, start_disksup, true),
{ok, _Pid} = supervisor:restart_child(os_mon_sup, disksup),
ok.
otp_5910(suite) -> [];
otp_5910(doc) ->
["Test that alarms are cleared if disksup crashes or "
"if OS_Mon is stopped"];
otp_5910(Config) when is_list(Config) ->
%% Make sure disksup sets at least one alarm
Data = disksup:get_disk_data(),
Threshold0 = disksup:get_almost_full_threshold(),
Threshold = case over_threshold(Data, Threshold0) of
0 ->
[{_Id,_Kbyte,Cap}|_] = Data,
ok = disksup:set_almost_full_threshold((Cap-1)/100),
Cap-1;
_N -> Threshold0
end,
ok = application:set_env(os_mon, disk_almost_full_threshold, Threshold/100),
disksup ! timeout, % force a disk check
Data2 = disksup:get_disk_data(),
Over = over_threshold(Data2, Threshold),
Alarms = get_alarms(),
if
Over==0 ->
?t:fail({threshold_too_low, Data2, Threshold});
Over==length(Alarms) ->
ok;
true ->
dump_info(),
?t:fail({bad_alarms, Threshold, Data2, Alarms})
end,
%% Kill disksup
exit(whereis(disksup), faked_disksup_crash),
%% Wait a little to make sure disksup has been restarted,
%% then make sure the alarms are set once, but not twice
?t:sleep(?t:seconds(1)),
Data3 = disksup:get_disk_data(),
Alarms2 = get_alarms(),
if
length(Alarms2)==length(Alarms) ->
ok;
true ->
dump_info(),
?t:fail({bad_alarms,Threshold,Data3,Alarms,Alarms2})
end,
%% Stop OS_Mon and make sure all disksup alarms are cleared
ok = application:stop(os_mon),
?t:sleep(?t:seconds(1)),
Alarms3 = get_alarms(),
case get_alarms() of
[] -> ok;
_ -> ?t:fail({alarms_not_cleared, Alarms3})
end,
%% Reset threshold and restart OS_Mon
ok = application:set_env(os_mon, disksup_almost_full_threshold, 0.8),
ok = disksup:set_almost_full_threshold(0.8),
ok = application:start(os_mon),
ok.
dump_info() ->
io:format("Status: ~p~n", [sys:get_status(disksup)]).
