- improved statfs system call
 - intercepted statvfs library call

Both of the above function now report the aggregated real size
of all the data-nodes.

You can test this with `df "/tmp/gekkofs_mountpoint"` command

Changes have been merged upstream [1] with commit `acb8c296a694d7e3e06756e7a102a9d1c92df0c8`.
There is no more need to patch margo to expose enanched init function.

[1]: https://xgitlab.cels.anl.gov/sds/margo/merge_requests/9

Paths passed as parameter to the daemon needs to be canonicalized

Since now there is only one margo instance there is no more difference
between RPC and IPC.

All the "ipc" occurence in the code have been removed in favor of "rpc"

the functionality wasn't used at all and it will be implemented later
with a different approach

Instead of passing only the mode as paramenter, we now pass an entire
Metadata object in such a way that the parameter can be configured at
compile time. For instace if we support symlinks we would like also to
pass the target path

Implemented all the logic to handle truncate operation.

Test: added truncate test

The ChunkStorage class now store the chunksize in order to perform
runtime checks upon chunks bundaries.

Mercury now support shared memory autorouting.

A single margo instance can be initialized and it will handle both
shared memory communication and remote ones.

If the endpoint of the RPC is local mercury will automatically use
shared memory.

Since there is only one margo instance all the duplicated code for
rpc/ipc have been unified and simplified considerably.

------

The way in which client contact the server has changed.

 - Server initializes its own margo instance and generate the endpoint
communication string using `HG_Addr_self`.
 - This endpoint description string is written on the pid file
 - When the client library loads it will fetch the server endpoint
description from the pid file and will use that to contact the server.

margo, argobots and mercury have been updated. The abt-snoozer dep
required by margo it is not needed anymore.

RocksDB has been also updated to the lastest release

loggers can be now configured at execution time, by using environment
variables for both server and client.

Server:
 - ADAFS_LOG_LEVEL
 - ADAFS_DAEMON_LOG_PATH

Client:
 - ADAFS_LOG_LEVEL
 - ADAFS_PRELOAD_LOG_PATH

Support for the following calls has been added:
 - opendir
 - readdir
 - closedir

The readdir call use the new RPC get_dirents that ask to all the
nodes (broadcast) about all the existing first-level entries
of a specific directory.

Now that we use only relative path on daemon side,
also the initial creation of the root directory should be
use a relative path.

Data operation has been grouped under a ChunkStorage class
that will manage all the data backend operations.

The instance of this object is stored into the ADAFS_DATA sigleton.

During daemon setup, multiple spdloggers are configured.

Using spdlog::get("<LOGNAME>") is possible to retrieve those and use
them in other part of the code.

Since the introduction of required parameters wasn't possible to show
the help message.
In fact the --help flag was checked after all other parameters.

Previously, clients would look for the daemon by name which was not robust.
If the daemon was started with valgrind for example, clients would not find
the daemon.

Similar to other applications a daemon now creates a pid file with
information about its pid and its started mountdir.

Proper errorhandling on server and client side was added.

Multiple daemons per node are explicitly not allowed for now!

Review and fixes that died when rebasing all commits

Nafi3 authored Mar 12, 2018 and Marc Vef committed Mar 23, 2018

This commit gives new options regarding separation of metadir and rootdir location. If only rootdir destination is given both metadata and data is stored at the same path

- Previously the directory hierarchy was not clear regarding to which
  file belong to which part of the project (client or daemon).
  Further, we will have other clients in the future (such as Fuse).
- CMake files now differentiate between include dirs for all targets
  and target specific ones.
- Removed duplicate -pg flag.
- Not listing header files when adding executables or libraries
  is considered bad practice. Note that include_directories() is adding
  include paths to the code while adding all files used for executables
  and libraries provide the context which files belong to each binary.
  When only include_directories() is set, CMake assumes that all files
  belong to a binary (which is not necessarily true). As a result,
  some IDEs may break as the do not support this assumption.

  In general we should almost always favor explicitness over implicitness.

ADA-FS's statfs will fill the struct mostly with rough estimations.
It is not supposed to return accurate data but "some" data for applications to work

Preload lib: RPC/IPC Mercury/Margo environments are only initialized if the mountdir path is used by the application
For that, the daemon writes the mountdir (it was started with) to a configurable path to disk.
The libraries read this value and only the process that uses part of this path will actually trigger the environment initialization.
This forbids prior behavior where applications initialize the whole environment numerous times, although unused.
For example, any MPI program will spawn multiple processes, all with the LD_PRELOAD environment variable set.