Implement couch_file:format_status to log filepath

[iilyak]

Overview

Erlang OTP logs the fact when gen_server behavior crashes. However in case of couch_file the filename is not part of the state. So it is quite hard to figure out what happened given the current log entry format. This PR adds information about filepath to the log entry. Keep in mind though that the filepath value is captured on couch_file:start_link. It is not representing the current file name if file is renamed or moved.

Testing recommendations

1. Start couchdb dev/run --admin=adm:pass
2. Start remsh session dev/remsh
3. Inject failure in one of the couch_file processes (by reading with position greater than the file size)

{_Port, Pid, _Fd, _Name} = hd(couch_debug:opened_files()).
couch_file:pread_binary(Pid, 999999999999999).

4. Check logs to make sure the following messages are present:

[error] 2018-09-12T20:01:02.957946Z node1@127.0.0.1 <0.390.0> -------- gen_server <0.390.0> terminated with reason: bad return value {read_beyond_eof,"/home/jenkins/dev/lib/node1/data/dbs/nodes.couch"}
  last msg: {pread_iolist,123123123}
     state: [{data,[{"State",{file,{file_descriptor,prim_file,{#Port<0.51359>,29}},true,139454,#Ref<0.0.3.65>,infinity}},{"InitialFilePath","/home/jenkins/dev/lib/node1/data/dbs/nodes.couch"}]}]

Related Issues or Pull Requests

This is a port of PR from apache/couchdb-couch#215 to couchdb repository.

Checklist

• Code is written and works correctly;
• Changes are covered by tests;
• Documentation reflects the changes;

[iilyak]

Reusing +1 from previous review of the same change apache/couchdb-couch#215 (comment)

[eiri]

This is a great fix, but why did you did it in CamelCase? This is so strange to see javaism in erlang's stack trace...

[iilyak]

@eiri: Would this work #1630 or you prefer a string "initial_file_path"?

[eiri]

@iilyak Personally I prefer atoms for both initial_file_path and state just so it's consistent with the rest of stack trace.

But to be clear I didn't try to twist your arm here, it's obviously a question of personal preferences, I was just was surprised when I was walking through commits happened while I was away.

[iilyak]

@eiri: Here is an extract from [gen_server documentation] (http://erlang.org/doc/man/gen_server.html#Module:format_status-2)

the recommended form for the Status value is [{data, [{"State", Term}]}], where Term provides relevant details of the gen_server state. Following this recommendation is not required, but it makes the callback module status consistent with the rest of the sys:get_status/1,2 return value.

I remember being inspired by the following example from OTP

format_status(_Opt, [_PDict, #state{os = OS, threshold = Threshold,
				    timeout = Timeout,
				    diskdata = DiskData}]) ->
    [{data, [{"OS", OS},
	     {"Timeout", Timeout},
	     {"Threshold", Threshold},
             {"DiskData", DiskData}]}].

The use of the function in otp is inconsistent:

• wx_object.erl

Specific = [{data, [{"State", State}]}],
[{header, Header},
     {data, [{"Status", SysState},
	     {"Parent", Parent},
	     {"Logged events", Log}]} |
Specfic]. 

• httpd_request_handler.erl

[{data, [{"StateData", State}]}]; 

• supervisor.erl

[{data, [{"State", State}]},
 {supervisor, [{"Callback", State#state.module}]}]

• memsup

[{data, [{"Timeout", Timeout}]},
 {items, {"Memory Usage", [
      {"Allocated", Allocated},
      {"Total", Total}]}},
 {items, {"Worst Memory User", WorstMemFormat}}].

• ssh_connection_handler.erl

[{data, [{"State", [{foo, 1}, {bar, 2}]}]}].

My guess is it might have to do with the way how observer or SASL uses this data. Since we are not using either we should be able to use any format we want.

[eiri]

Huh. I haven't seen this gen_server doc's snippet, thank you for pointing out.

To be honest I feel very confused on what OTP folks are doing here in source code, mixing strings, atoms, camelcase and capitalized spaced. Dunno, maybe it just me.

Anyway, I take my complaints back, since you are following worded OTP recommendation here let's keep it this way, sorry about whole confusion.