detect bad merge index files

gistfile1.erl

#!/usr/bin/env escript
%% -*- erlang -*-

-include_lib("kernel/include/file.hrl").
-compile(export_all).
-define(LOG(S), io:format(S)).
-define(LOG(S,A), io:format(S,A)).

main(Dirs) ->
    CodePath = case os:getenv("RIAK_LIB") of
                   false -> throw("RIAK_LIB must be set");
                   Val -> filelib:wildcard(filename:join([Val, "*/ebin"]))
               end,
    code:add_paths(CodePath),
    code:ensure_loaded(mi_segment),
    case code:is_loaded(mi_segment) of
        false -> throw("Could not load mi_segment");
        _ -> ok
    end,
    application:load(merge_index),
    [detect_bad_files(D) || D <- Dirs];
main(_) ->
    usage().

usage() ->
    io:format("usage: detect-bad-files <code-path> <mi-root-dir>~n").

detect_bad_files(Dir) ->
    detect_bad_buffers(Dir),
    detect_unmatched_segs(Dir),
    detect_bad_seg_offsets(Dir),
    detect_bad_segs(Dir).

detect_bad_buffers(Dir) ->
    Buffs = file_glob(Dir, "buffer*"),
    [check_buffer(B) || B <- Buffs].

check_buffer(Buff) ->
    try
        mi_buffer:new(Buff)
    catch _:_ ->
        ?LOG("~p~n", [Buff])
    end.

detect_unmatched_segs(Dir) ->
    Offsets = file_glob(Dir, "*.offsets"),
    Segs = file_glob(Dir, "*.data"),
    Offsets2 = ordsets:from_list([offset_clean(O) || O <- Offsets]),
    Segs2 = ordsets:from_list([seg_clean(S) || S <- Segs]),
    MissingSegs = ordsets:subtract(Offsets2, Segs2),
    MissingOffsets = ordsets:subtract(Segs2, Offsets2),
    [missing_seg(S) || S <-  MissingSegs],
    [missing_offset(O) || O <- MissingOffsets].

offset_clean(OffsetFile) ->
    filename:basename(OffsetFile, ".offsets").

seg_clean(SegFile) ->
    filename:basename(SegFile, ".data").

missing_seg(Seg) ->
    ?LOG("MISSING_SEGMENT: ~p~n", [Seg]).

missing_offset(Offset) ->
    ?LOG("MISSING_SEGMENT_OFFSET: ~p~n", [Offset]).

detect_bad_seg_offsets(Dir) ->
    Offsets = file_glob(Dir, "*.offsets"),
    [check_offset(O) || O <- Offsets].

check_offset(Offset) ->
    try
        ets:file2tab(Offset)
    catch _ ->
            ?LOG("~p~n",
                 [Offset])
    end.

detect_bad_segs(Dir) ->
    Segs = file_glob(Dir, "*.data"),
    [check_seg(S) || S <- Segs].

check_seg(Seg) ->
    try
        {ok, FileInfo} = file:read_file_info(Seg),
        Seg2 = {segment, filename:rootname(Seg), dummy_offset_tab, FileInfo#file_info.size},
        Itr = mi_segment:iterator(Seg2),
        iterate(Itr())
    catch _:_ ->
            ?LOG("~p~n", [Seg])
    end.

iterate(eof) ->
    ok;
iterate({_,Itr}) ->
    iterate(Itr()).

file_glob(Dir, Glob) ->
    filelib:wildcard(filename:join([Dir, Glob])).