When reading loose objects over NFS it is possible that the OS syscall
would fail with ESTALE errors: This happens when the open file
descriptor no longer refers to a valid file.

Notoriously it is possible to hit this scenario when git data is shared
among multiple clients, for example by multiple gerrit instances in HA.

If one of the two clients performs a GC operation that would cause the
packing and then the pruning of loose objects, the other client might
still hold a reference to those objects, which would cause an exception
to bubble up the stack.

The Linux NFS FAQ[1] (at point A.10), suggests that the proper way to
handle such ESTALE scenarios is to:

"[...] close the file or directory where the error occurred, and reopen
it so the NFS client can resolve the pathname again and retrieve the new
file handle."

In case of a stale file handle exception, we now attempt to read the
loose object again (up to 5 times), until we either succeed or encounter
a FileNotFoundException, in which case the search can continue to
Packfiles and alternates.

The limit of 5 provides an arbitrary upper bounds that is consistent to
the one chosen when handling stale file handles for packed-refs
files (see [2] for context).

[1] http://nfs.sourceforge.net/
[2] https://git.eclipse.org/r/c/jgit/jgit/+/54350

Bug: 573791
Change-Id: I9950002f772bbd8afeb9c6108391923be9d0ef51

Change-Id: I64b0c057dd0a12aef2f3d56fa0c8a10e3b23fffd
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>

This enables higher file timestamp resolution on filesystems like ext4,
Mac APFS (1ns) or NTFS (100ns) providing high timestamp resolution on
filesystem level.

Note:
- on some OSes Java 8,9 truncate milliseconds, see
https://bugs.openjdk.java.net/browse/JDK-8177809, fixed in Java 10
- UnixFileAttributes truncates timestamp resolution to microseconds when
converting the internal representation to FileTime exposed in the API,
see https://bugs.openjdk.java.net/browse/JDK-8181493
- WindowsFileAttributes also provides only microsecond resolution

Change-Id: I25ffff31a3c6f725fc345d4ddc2f26da3b88f6f2
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>

Convert anonymous inner classes to lambda expressions or member
references

Bug: 545856
CQ: 19537
Change-Id: I621431c178e8b99316314602f7c66c9a36f9ae98
Signed-off-by: Carsten Hammer <carsten.hammer@t-online.de>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>

Do not reload packfiles when their associated filesnapshot is not
modified on disk compared to the one currently stored in memory.

Fix the regression introduced by fef78212 which, in conjunction with
core.trustfolderstats = false, caused any lookup of objects inside
the packlist to loop forever when the object was not found in the pack
list.

Bug: 546190
Change-Id: I38d752ebe47cefc3299740aeba319a2641f19391
Signed-off-by: Luca Milanesio <luca.milanesio@gmail.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>

Its implementation contains

		} catch (IOException e) {
			// Legacy API, assume error means "no"
			return false;
		}

Better to use ObjectDatabase#has, which throws IOException to report
errors.

Change-Id: I7de02f7ceb8f57b2a8ebdb16d2aa4376775ff933
Signed-off-by: Jonathan Nieder <jrn@google.com>

Change-Id: Ie1631784b5eba04edb21f66df284f9e279fd2ec0
Signed-off-by: David Pursehouse <david.pursehouse@gmail.com>

So far, in order to get the pack directory it was necessary to resolve
it from the object directory. This resolution is already done when
creating the object directory, so simplify the call by just adding a
getter to the pack directory.

Change-Id: I69e783141dc6739024e8b3d5acc30843edd651a7
Signed-off-by: Hector Caballero <hector.caballero@ericsson.com>

Set missingOverrideAnnotation=warning in Eclipse compiler preferences
which enables the warning:

  The method <method> of type <type> should be tagged with @Override
  since it actually overrides a superclass method

Justification for this warning is described in:

  http://stackoverflow.com/a/94411/381622

Enabling this causes in excess of 1000 warnings across the entire
code-base. They are very easy to fix automatically with Eclipse's
"Quick Fix" tool.

Fix all of them except 2 which cause compilation failure when the
project is built with mvn; add TODO comments on those for further
investigation.

Change-Id: I5772061041fd361fe93137fd8b0ad356e748a29c
Signed-off-by: David Pursehouse <david.pursehouse@gmail.com>

ObjectDirectory.getShallowCommits should throw an IOException
instead of an InvalidArgumentException if invalid SHAs are present
in .git/shallow (as this file is usually edited by a human).

Change-Id: Ia3a39d38f7aec4282109c7698438f0795fbec905
Signed-off-by: Marc Strapetz <marc.strapetz@syntevo.com>
Signed-off-by: David Pursehouse <david.pursehouse@gmail.com>

Change-Id: I0307d1e9bbc0bfc8172eed10f8e93788d6e2c8b4
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>

FileSnapshot.isModified may have reported a file to be clean although it
was actually dirty.

Imagine you have a FileSnapshot on file f. lastmodified and lastread are
both t0. Now time is t1 and you
1) modify the file
2) update the FileSnapshot of the file (lastModified=t1, lastRead=t1)
3) modify the file again
4) wait 3 seconds
5) ask the Filesnapshot whether the file is dirty or not. It erroneously
answered it's clean.

Any file which has been modified longer than 2.5 seconds ago was
reported to be clean. As the test shows that's not always correct.

The real-world problem fixed by this change is the following:
* A gerrit server using JGit to serve git repositories is processing
fetch requests while simultaneously a native git garbage collection
runs on the repo.
* At time t1 native git writes temporary files in the pack folder
setting the mtime of the pack folder to t1.
* A fetch request causes JGit to search for new packfiles and JGit
remembers this scan in a Filesnapshot on the packs folder. Since the gc
is not finished JGit doesn't see any new packfiles.
* The fetch is processed and the gc ends while the filesystem timer is
still t1. GC writes a new packfile and deletes the old packfile.
* 3 seconds later another request arrives. JGit does not yet know about
the new packfile but is also not rescanning the pack folder because it
cached that the last scan happened at time t1 and pack folder's mtime is
also t1. Now JGit will not be able to resolve any object contained in
this new pack. This behavior may be persistent if objects referenced by
the ref/meta/config branch are affected so gerrit can't read permissions
stored in the refs/meta/config branch anymore and will not allow any
pushes anymore. The pack folder will not change its mtime and therefore
no rescan will take place.

Change-Id: I3efd0ccffeb97b01207dc3e7a6b85c6b06928fad
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>

Change-Id: I7c545d06b1bced678c020fab9af1382bc4416b6e
Signed-off-by: David Pursehouse <david.pursehouse@gmail.com>

The variable and parameter named 'db' were hiding class members
with the same name.

Change-Id: I27017afdc5f49c38c6f5be494e7a21239ea601a7
Signed-off-by: David Pursehouse <david.pursehouse@sonymobile.com>

This breaks all existing callers once. Applications are not supposed
to build against the internal storage API unless they can accept API
churn and make necessary updates as versions change.

Change-Id: I2ab1327c202ef2003565e1b0770a583970e432e9

Change-Id: Iedb54eb9d8396bc3ae66d8754c1527fd9ca655f9

If multiple threads attempted to insert loose objects into the same new
fan-out directory, the creation of that directory was subject to a race
condition that could lead to an unnecessary IOException being thrown -
because an inserter could not 'create' a directory that had just been
generated by a different thread. All we require is that the directory
does indeed *exist*, so not being able to _create_ it is not actually a
fatal problem. Setting 'skipExisting' to 'true' on the call to mkdir()
fixes the issue.

I found this issue as a real world occurrence while working on The BFG
Repo Cleaner (https://github.com/rtyley/bfg-repo-cleaner), a tool which
concurrently performs a lot of object creation.

In order to demonstrate the problem here I've added a small test case
which reliably reproduces the issue on the few different hardware
systems I've tried. The error thrown when the race-condition arises is
this:

java.io.IOException: Creating directory /home/roberto/repo.git/objects/e6 failed
at org.eclipse.jgit.util.FileUtils.mkdir(FileUtils.java:182)
at org.eclipse.jgit.storage.file.ObjectDirectory.insertUnpackedObject(ObjectDirectory.java:590)
at org.eclipse.jgit.storage.file.ObjectDirectoryInserter.insertOneObject(ObjectDirectoryInserter.java:113)
at org.eclipse.jgit.storage.file.ObjectDirectoryInserter.insert(ObjectDirectoryInserter.java:91)
at org.eclipse.jgit.lib.ObjectInserter.insert(ObjectInserter.java:329)

Change-Id: I88eac49bc600c56ba9ad290e6133d8a7113125ab

In practice the DHT storage layer has not been performing as well as
large scale server environments want to see from a Git server.

The performance of the DHT schema degrades rapidly as small changes
are pushed into the repository due to the chunk size being less than
1/3 of the pushed pack size.  Small chunks cause poor prefetch
performance during reading, and require significantly longer prefetch
lists inside of the chunk meta field to work around the small size.

The DHT code is very complex (>17,000 lines of code) and is very
sensitive to the underlying database round-trip time, as well as the
way objects were written into the pack stream that was chunked and
stored on the database.  A poor pack layout (from any version of C Git
prior to Junio reworking it) can cause the DHT code to be unable to
enumerate the objects of the linux-2.6 repository in a completable
time scale.

Performing a clone from a DHT stored repository of 2 million objects
takes 2 million row lookups in the DHT to locate the OBJECT_INDEX row
for each object being cloned. This is very difficult for some DHTs to
scale, even at 5000 rows/second the lookup stage alone takes 6 minutes
(on local filesystem, this is almost too fast to bother measuring).
Some servers like Apache Cassandra just fall over and cannot complete
the 2 million lookups in rapid fire.

On a ~400 MiB repository, the DHT schema has an extra 25 MiB of
redundant data that gets downloaded to the JGit process, and that is
before you consider the cost of the OBJECT_INDEX table also being
fully loaded, which is at least 223 MiB of data for the linux kernel
repository.  In the DHT schema answering a `git clone` of the ~400 MiB
linux kernel needs to load 248 MiB of "index" data from the DHT, in
addition to the ~400 MiB of pack data that gets sent to the client.
This is 193 MiB more data to be accessed than the native filesystem
format, but it needs to come over a much smaller pipe (local Ethernet
typically) than the local SATA disk drive.

I also never got around to writing the "repack" support for the DHT
schema, as it turns out to be fairly complex to safely repack data in
the repository while also trying to minimize the amount of changes
made to the database, due to very common limitations on database
mutation rates..

This new DFS storage layer fixes a lot of those issues by taking the
simple approach for storing relatively standard Git pack and index
files on an abstract filesystem. Packs are accessed by an in-process
buffer cache, similar to the WindowCache used by the local filesystem
storage layer. Unlike the local file IO, there are some assumptions
that the storage system has relatively high latency and no concept of
"file handles". Instead it looks at the file more like HTTP byte range
requests, where a read channel is a simply a thunk to trigger a read
request over the network.

The DFS code in this change is still abstract, it does not store on
any particular filesystem, but is fairly well suited to the Amazon S3
or Apache Hadoop HDFS. Storing packs directly on HDFS rather than
HBase removes a layer of abstraction, as most HBase row reads turn
into an HDFS read.

Most of the DFS code in this change was blatently copied from the
local filesystem code. Most parts should be refactored to be shared
between the two storage systems, but right now I am hesistent to do
this due to how well tuned the local filesystem code currently is.

Change-Id: Iec524abdf172e9ec5485d6c88ca6512cd8a6eafb

jgit.storage.dht is a storage provider implementation for JGit that
permits storing the Git repository in a distributed hashtable, NoSQL
system, or other database.  The actual underlying storage system is
undefined, and can be plugged in by implementing 7 small interfaces:

  *  Database
  *  RepositoryIndexTable
  *  RepositoryTable
  *  RefTable
  *  ChunkTable
  *  ObjectIndexTable
  *  WriteBuffer

The storage provider interface tries to assume very little about the
underlying storage system, and requires only three key features:

  *  key -> value lookup (a hashtable is suitable)
  *  atomic updates on single rows
  *  asynchronous operations (Java's ExecutorService is easy to use)

Most NoSQL database products offer all 3 of these features in their
clients, and so does any decent network based cache system like the
open source memcache product.  Relying only on key equality for data
retrevial makes it simple for the storage engine to distribute across
multiple machines.  Traditional SQL systems could also be used with a
JDBC based spi implementation.

Before submitting this change I have implemented six storage systems
for the spi layer:

  * Apache HBase[1]
  * Apache Cassandra[2]
  * Google Bigtable[3]
  * an in-memory implementation for unit testing
  * a JDBC implementation for SQL
  * a generic cache provider that can ride on top of memcache

All six systems came in with an spi layer around 1000 lines of code to
implement the above 7 interfaces.  This is a huge reduction in size
compared to prior attempts to implement a new JGit storage layer.  As
this package shows, a complete JGit storage implementation is more
than 17,000 lines of fairly complex code.

A simple cache is provided in storage.dht.spi.cache.  Implementers can
use CacheDatabase to wrap any other type of Database and perform fast
reads against a network based cache service, such as the open source
memcached[4].  An implementation of CacheService must be provided to
glue this spi onto the network cache.

[1] https://github.com/spearce/jgit_hbase
[2] https://github.com/spearce/jgit_cassandra
[3] http://labs.google.com/papers/bigtable.html
[4] http://memcached.org/

Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>

There is no point in pushing all of the files within the edge
commits into the delta search when making a thin pack.  This floods
the delta search window with objects that are unlikely to be useful
bases for the objects that will be written out, resulting in lower
data compression and higher transfer sizes.

Instead observe the path of a tree or blob that is being pushed
into the outgoing set, and use that path to locate up to WINDOW
ancestor versions from the edge commits.  Push only those objects
into the edgeObjects set, reducing the number of objects seen by the
search window.  This allows PackWriter to only look at ancestors
for the modified files, rather than all files in the project.
Limiting the search to WINDOW size makes sense, because more than
WINDOW edge objects will just skip through the window search as
none of them need to be delta compressed.

To further improve compression, sort edge objects into the front
of the window list, rather than randomly throughout.  This puts
non-edges later in the window and gives them a better chance at
finding their base, since they search backwards through the window.

These changes make a significant difference in the thin-pack:

  Before:
    remote: Counting objects: 144190, done
    remote: Finding sources: 100% (50275/50275)
    remote: Getting sizes: 100% (101405/101405)
    remote: Compressing objects: 100% (7587/7587)
    Receiving objects: 100% (50275/50275), 24.67 MiB | 9.90 MiB/s, done.
    Resolving deltas: 100% (40339/40339), completed with 2218 local objects.

    real    0m30.267s

  After:
    remote: Counting objects: 61549, done
    remote: Finding sources: 100% (50275/50275)
    remote: Getting sizes: 100% (18862/18862)
    remote: Compressing objects: 100% (7588/7588)
    Receiving objects: 100% (50275/50275), 11.04 MiB | 3.51 MiB/s, done.
    Resolving deltas: 100% (43160/43160), completed with 5014 local objects.

    real    0m22.170s

The resulting pack is 13.63 MiB smaller, even though it contains the
same exact objects.  82,543 fewer objects had to have their sizes
looked up, which saved about 8s of server CPU time.  2,796 more
objects from the client were used as part of the base object set,
which contributed to the smaller transfer size.

Change-Id: Id01271950432c6960897495b09deab70e33993a9
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Sigend-off-by: Chris Aniszczyk <caniszczyk@gmail.com>

Per CQ 3448 this is the initial contribution of the JGit project
to eclipse.org.  It is derived from the historical JGit repository
at commit 3a2dd9921c.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>