[PATCH 01/14] graph: add packed graph design document

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Add Documentation/technical/packed-graph.txt with details of the planned
packed graph feature, including future plans.

Signed-off-by: Derrick Stolee <dstolee@xxxxxxxxxxxxx>
---
 Documentation/technical/packed-graph.txt | 185 +++++++++++++++++++++++++++++++
 1 file changed, 185 insertions(+)
 create mode 100644 Documentation/technical/packed-graph.txt

diff --git a/Documentation/technical/packed-graph.txt b/Documentation/technical/packed-graph.txt
new file mode 100644
index 0000000000..fcc0c83874
--- /dev/null
+++ b/Documentation/technical/packed-graph.txt
@@ -0,0 +1,185 @@
+Git Packed Graph Design Notes
+=============================
+
+Git walks the commit graph for many reasons, including:
+
+1. Listing and filtering commit history.
+2. Computing merge bases.
+
+These operations can become slow as the commit count grows above 100K.
+The merge base calculation shows up in many user-facing commands, such
+as 'status' and 'fetch' and can take minutes to compute depending on
+data shape. There are two main costs here:
+
+1. Decompressing and parsing commits.
+2. Walking the entire graph to avoid topological order mistakes.
+
+The packed graph is a file that stores the commit graph structure along
+with some extra metadata to speed up graph walks. This format allows a
+consumer to load the following info for a commit:
+
+1. The commit OID.
+2. The list of parents.
+3. The commit date.
+4. The root tree OID.
+5. An integer ID for fast lookups in the graph.
+6. The generation number (see definition below).
+
+Values 1-4 satisfy the requirements of parse_commit_gently().
+
+By providing an integer ID we can avoid lookups in the graph as we walk
+commits. Specifically, we need to provide the integer ID of the parent
+commits so we navigate directly to their information on request.
+
+Define the "generation number" of a commit recursively as follows:
+ * A commit with no parents (a root commit) has generation number 1.
+ * A commit with at least one parent has generation number 1 more than
+   the largest generation number among its parents.
+Equivalently, the generation number is one more than the length of a
+longest path from the commit to a root commit. The recursive definition
+is easier to use for computation and the following property:
+
+    If A and B are commits with generation numbers N and M, respectively,
+    and N <= M, then A cannot reach B. That is, we know without searching
+    that B is not an ancestor of A because it is further from a root commit
+    than A.
+
+    Conversely, when checking if A is an ancestor of B, then we only need
+    to walk commits until all commits on the walk boundary have generation
+    number at most N. If we walk commits using a priority queue seeded by
+    generation numbers, then we always expand the boundary commit with highest
+    generation number and can easily detect the stopping condition.
+
+This property can be used to significantly reduce the time it takes to
+walk commits and determine topological relationships. Without generation
+numbers, the general heuristic is the following:
+
+    If A and B are commits with commit time X and Y, respectively, and
+    X < Y, then A _probably_ cannot reach B.
+
+This heuristic is currently used whenever the computation can make
+mistakes with topological orders (such as "git log" with default order),
+but is not used when the topological order is required (such as merge
+base calculations, "git log --graph").
+
+Design Details
+--------------
+
+- A graph file is stored in a file named 'graph-<oid>.graph' in the pack
+  directory. This could be stored in an alternate.
+
+- The most-recent graph file OID is stored in a 'graph-head' file for
+  immediate access and storing backup graphs. This could be stored in an
+  alternate, and refers to a 'graph-<oid>.graph' file in the same pack
+  directory.
+
+- The core.graph config setting must be on to create or consume graph files.
+
+- The graph file is only a supplemental structure. If a user downgrades
+  or disables the 'core.graph' config setting, then the existing ODB is
+  sufficient.
+
+- The file format includes parameters for the object id length
+  and hash algorithm, so a future change of hash algorithm does
+  not require a change in format.
+
+Current Limitations
+-------------------
+
+- Only one graph file is used at one time. This allows the integer ID to
+  seek into the single graph file. It is possible to extend the model
+  for multiple graph files, but that is currently not part of the design.
+
+- .graph files are managed only by the 'graph' builtin. These are not
+  updated automatically during clone or fetch.
+
+- There is no '--verify' option for the 'graph' builtin to verify the
+  contents of the graph file.
+
+- The graph only considers commits existing in packfiles and does not
+  walk to fill in reachable commits. [Small]
+
+- When rewriting the graph, we do not check for a commit still existing
+  in the ODB, so garbage collection may remove commits
+
+- Generation numbers are not computed in the current version. The file
+  format supports storing them, along with a mechanism to upgrade from
+  a file without generation numbers to one that uses them.
+
+Future Work
+-----------
+
+- The file format includes room for precomputed generation numbers. These
+  are not currently computed, so all generation numbers will be marked as
+  0 (or "uncomputed"). A later patch will include this calculation.
+
+- The current implementation of the 'graph' builtin walks all packed objects
+  to find a complete list of commits in packfiles. If some commits are
+  stored as loose objects, then these do not appear in the graph. This is
+  handled gracefully by the file format, but it would cause incorrect
+  generation number calculations. We should implement the construct_graph()
+  method in a way that walks all commits reachable from some starting set
+  and then can use complete information for generation numbers. (Some
+  care must be taken around shallow clones.)
+
+- The graph is not currently integrated with grafts.
+
+- After computing and storing generation numbers, we must make graph
+  walks aware of generation numbers to gain performance benefits. This
+  will mostly be accomplished by swapping a commit-date-ordered priority
+  queue with one ordered by generation number. The following operations
+  are important candidates:
+
+    - paint_down_to_common()
+    - 'log --topo-order'
+
+- The graph currently only adds commits to a previously existing graph.
+  When writing a new graph, we could check that the ODB still contains
+  the commits and choose to remove the commits that are deleted from the
+  ODB. For performance reasons, this check should remain optional.
+
+- Currently, parse_commit_gently() requires filling in the root tree
+  object for a commit. This passes through lookup_tree() and consequently
+  lookup_object(). Also, it calls lookup_commit() when loading the parents.
+  These method calls check the ODB for object existence, even if the
+  consumer does not need the content. For example, we do not need the
+  tree contents when computing merge bases. Now that commit parsing is
+  removed from the computation time, these lookup operations are the
+  slowest operations keeping graph walks from being fast. Consider
+  loading these objects without verifying their existence in the ODB and
+  only loading them fully when consumers need them. Consider a method
+  such as "ensure_tree_loaded(commit)" that fully loads a tree before
+  using commit->tree.
+
+- The current design uses the 'graph' builtin to generate the graph. When
+  this feature stabilizes enough to recommend to most users, we should
+  add automatic graph writes to common operations that create many commits.
+  For example, one coulde compute a graph on 'clone' and 'fetch' commands.
+
+Related Links
+-------------
+
+[0] https://bugs.chromium.org/p/git/issues/detail?id=8
+    Chromium work item for: Serialized Commit Graph
+
+[1] https://public-inbox.org/git/20110713070517.GC18566@xxxxxxxxxxxxxxxxxxxxx/
+    An abandoned patch that introduced generation numbers.
+
+[2] https://public-inbox.org/git/20170908033403.q7e6dj7benasrjes@xxxxxxxxxxxxxxxxxxxxx/
+    Discussion about generation numbers on commits and how they interact
+    with fsck.
+
+[3] https://public-inbox.org/git/20170907094718.b6kuzp2uhvkmwcso@xxxxxxxxxxxxxxxxxxxxx/t/#m7a2ea7b355aeda962e6b86404bcbadc648abfbba
+    More discussion about generation numbers and not storing them inside
+    commit objects. A valuable quote:
+
+    "I think we should be moving more in the direction of keeping
+     repo-local caches for optimizations. Reachability bitmaps have been
+     a big performance win. I think we should be doing the same with our
+     properties of commits. Not just generation numbers, but making it
+     cheap to access the graph structure without zlib-inflating whole
+     commit objects (i.e., packv4 or something like the "metapacks" I
+     proposed a few years ago)."
+
+[4] https://public-inbox.org/git/20180108154822.54829-1-git@xxxxxxxxxxxxxxxxx/T/#u
+    A patch to remove the ahead-behind calculation from 'status'.
-- 
2.16.0




[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]

  Powered by Linux