On 6/20/2017 3:54 AM, Christian Couder wrote:
Goal
~~~~
Git can store its objects only in the form of loose objects in
separate files or packed objects in a pack file.
To be able to better handle some kind of objects, for example big
blobs, it would be nice if Git could store its objects in other object
databases (ODB).
To do that, this patch series makes it possible to register commands,
also called "helpers", using "odb.<odbname>.command" config variables,
to access external ODBs where objects can be stored and retrieved.
External ODBs should be able to tranfer information about the blobs
they store. This patch series shows how this is possible using kind of
replace refs.
Great to see this making progress!
My thoughts and questions are mostly about the overall design tradeoffs.
Is your intention to enable the ODB to completely replace the regular
object store or just to supplement it? I think it would be good to
ensure the interface is robust and performant enough to actually replace
the current object store interface (even if we don't actually do that
just yet).
Another way of asking this is: do the 3 verbs (have, get, put) and the 3
types of "get" enable you to wrap the current loose object and pack file
code as ODBs and run completely via the external ODB interface? If not,
what is missing and can it be added?
_Eventually_ it would be great to see the current object store(s) moved
behind the new ODB interface.
When there are multiple ODB providers, what is the order they are
called? If one fails a request (get, have, put) are the others called
to see if they can fulfill the request?
Can the order they are called for various verb be configured explicitly?
For example, it would be nice to have a "large object ODB handler"
configured to get first try at all "put" verbs. Then if it meets it's
size requirements, it will handle the verb, otherwise it fail and git
will try the other ODBs.
Design
~~~~~~
* The "helpers" (registered commands)
Each helper manages access to one external ODB.
There are now 2 different modes for helper:
- When "odb.<odbname>.scriptMode" is set to "true", the helper is
launched each time Git wants to communicate with the <odbname>
external ODB.
- When "odb.<odbname>.scriptMode" is not set or set to "false", then
the helper is launched once as a sub-process (using
sub-process.h), and Git communicates with it using packet lines.
Is it worth supporting two different modes long term? It seems that
this could be simplified (less code to write, debug, document, support)
by only supporting the 2nd that uses the sub-process. As far as I can
tell, the capabilities are the same, it's just the second one is more
performant when multiple calls are made.
A helper can be given different instructions by Git. The instructions
that are supported are negociated at the beginning of the
communication using a capability mechanism.
For now the following instructions are supported:
- "have": the helper should respond with the sha1, size and type of
all the objects the external ODB contains, one object per line.
- "get <sha1>": the helper should then read from the external ODB
the content of the object corresponding to <sha1> and pass it to Git.
- "put <sha1> <size> <type>": the helper should then read from from
Git an object and store it in the external ODB.
Currently "have" and "put" are optional.
It's good the various verbs can be optional. That way any particular
ODB only has to handle those it needs to provide a different behavior for.
There are 3 different kinds of "get" instructions depending on how the
helper passes objects to Git:
- "fault_in": the helper will write the requested objects directly
into the regular Git object database, and then Git will retry
reading it from there.
I think the "fault_in" behavior can be implemented efficiently without
the overhead of a 3rd special "get" instruction if we enable some of the
other capabilities discussed.
For example, assume an ODB is setup to handle missing objects (by
registering itself as "last" in the prioritized list of ODB handlers).
If it is ever asked to retrieve a missing object, it can retrieve the
object and return it as a "git_object" or "plain_object" and also cache
it locally as a loose object, pack file, or any other ODB handler
supported mechanism. Future requests will then provide that object via
the locally cached copy and its associated ODB handler.
- "git_object": the helper will send the object as a Git object.
- "plain_object": the helper will send the object (a blob) as a raw
object. (The blob content will be sent as is.)
For now the kind of "get" that is supported is read from the
"odb.<odbname>.fetchKind" configuration variable, but in the future it
should be decided as part of the capability negociation.
I agree it makes sense to move this into the capability negotiation but
I also wonder if we really need to support both. Is there a reason we
can't just choose one and force all ODBs to support it?
* Transfering information
This whole section on "odb ref" feels out of place to me. Below you
state it is optional in which case I think it should be discussed in the
patches that implement the tests that use it rather than here. It seems
to be a test ODB specific implementation detail.
To tranfer information about the blobs stored in external ODB, some
special refs, called "odb ref", similar as replace refs, are used in
the tests of this series, but in general nothing forces the helper to
use that mechanism.
The external odb helper is responsible for using and creating the refs
in refs/odbs/<odbname>/, if it wants to do that. It is free for example
to just create one ref, as it is also free to create many refs. Git
would just transmit the refs that have been created by this helper, if
Git is asked to do so.
For now in the tests there is one odb ref per blob, as it is simple
and as it is similar to what git-lfs does. Each ref name is
refs/odbs/<odbname>/<sha1> where <sha1> is the sha1 of the blob stored
in the external odb named <odbname>.
These odb refs point to a blob that is stored in the Git
repository and contain information about the blob stored in the
external odb. This information can be specific to the external odb.
The repos can then share this information using commands like:
`git fetch origin "refs/odbs/<odbname>/*:refs/odbs/<odbname>/*"`
At the end of the current patch series, "git clone" is teached a
"--initial-refspec" option, that asks it to first fetch some specified
refs. This is used in the tests to fetch the odb refs first.
> This way only one "git clone" command can setup a repo using the
external ODB mechanism as long as the right helper is installed on the
machine and as long as the following options are used:
- "--initial-refspec <odbrefspec>" to fetch the odb refspec
- "-c odb.<odbname>.command=<helper>" to configure the helper
There is also a test script that shows that the "--initial-refspec"
option along with the external ODB mechanism can be used to implement
cloning using bundles.
The fact that "git clone is taught a --initial-refspec" option"
indicates this isn't just an ODB implementation detail. Is there a
general capability that is missing from the ODB interface that needs to
be addressed here?
I don't believe there is. Instead, I think we should allow the various
"partial clone" patch series already in progress solve the problem of
how you do a partial clone of a repo.
[1]
https://public-inbox.org/git/1488994685-37403-1-git-send-email-jeffhost@xxxxxxxxxxxxx/
[2] https://public-inbox.org/git/20170309073117.g3br5btsfwntcdpe@xxxxxxxxxxxxxxxxxxxxx/
[3] https://public-inbox.org/git/cover.1496361873.git.jonathantanmy@xxxxxxxxxx/
[4] https://public-inbox.org/git/20170602232508.GA21733@xxxxxxxxxxxxxxxxxxxxxxxxx/
* External object database
This RFC patch series shows in the tests:
- how to use another git repository as an external ODB (storing Git objects)
- how to use an http server as an external ODB (storing plain objects)
(This works in both script mode and sub-process mode.)
* Performance
So the sub-process mode, which is now the default, has been
implemented in this new version of this patch series.
This has been implemented using the refactoring that Ben Peart did on
top of Lars Schneider's work on using sub-processes and packet lines
in the smudge/clean filters for git-lfs. This also uses further work
from Ben Peart called "read object process".
See:
http://public-inbox.org/git/20170113155253.1644-1-benpeart@xxxxxxxxxxxxx/
http://public-inbox.org/git/20170322165220.5660-1-benpeart@xxxxxxxxxxxxx/
Thanks to this, the external ODB mechanism should in the end perform
as well as the git-lfs mechanism when many objects should be
transfered.
Implementation
~~~~~~~~~~~~~~
* Mechanism to call the registered commands
This series adds a set of function in external-odb.{c,h} that are
called by the rest of Git to manage all the external ODBs.
These functions use 'struct odb_helper' and its associated functions
defined in odb-helper.{c,h} to talk to the different external ODBs by
launching the configured "odb.<odbname>.command" commands and writing
to or reading from them.
* ODB refs
For now odb ref management is only implemented in a helper in t0410.
When a new blob is added to an external odb, its sha1, size and type
are writen in another new blob and the odb ref is created.
When the list of existing blobs is requested from the external odb,
the content of the blobs pointed to by the odb refs can also be used
by the odb to claim that it can get the objects.
When a blob is actually requested from the external odb, it can use
the content stored in the blobs pointed to by the odb refs to get the
actual blobs and then pass them.
Highlevel view of the patches in the series
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Patch 1/49 is a small code cleanup that I already sent to the
mailing list but will probably be removed in the end bdue to
ongoing work on "git clone"
- Patches 02/49 to 08/49 create a Git/Packet.pm module by
refactoring "t0021/rot13-filter.pl". Functions from this new
module will be used later in test scripts.
Nice!
- Patches 09/49 to 16/49 create the external ODB insfrastructure
in external-odb.{c,h} and odb-helper.{c,h} for the script mode.
- Patches 17/49 to 23/49 improve lib-http to make it possible to
use it as an external ODB to test storing blobs in an HTTP
server.
- Patches 24/49 to 44/49 improve the external ODB insfrastructure
to support sub-processes and make everything work using them.
I understand why it is this way historically but it seems these should
just be combined with patches 9-16. Instead of writing the new odb
specific routines and then patching them to support sub-process, just
write them that way the "first" time.
- Patches 45/49 to 49/49 add the --initial-refspec to git clone
along with tests.
I'm hopeful the changes to git clone to add the --initial-refspec can
eventually be dropped. It seems we should be able to test the ODB
capabilities without having to add options to git clone.
Future work
~~~~~~~~~~~
First sorry about the state of this patch series, it is not as clean
as I would have liked, butI think it is interesting to get feedback
from the mailing list at this point, because the previous RFC was sent
a long time ago and a lot of things changed.
So a big part of the future work will be about cleaning this patch series.
Other things I think I am going to do:
-
Previous work and discussions
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
(Sorry for the old Gmane links, I will try to replace them with
public-inbox.org at one point.)
Peff started to work on this and discuss this some years ago:
http://thread.gmane.org/gmane.comp.version-control.git/206886/focus=207040
http://thread.gmane.org/gmane.comp.version-control.git/247171
http://thread.gmane.org/gmane.comp.version-control.git/202902/focus=203020
His work, which is not compile-tested any more, is still there:
https://github.com/peff/git/commits/jk/external-odb-wip
Initial discussions about this new series are there:
http://thread.gmane.org/gmane.comp.version-control.git/288151/focus=295160
Version 1, 2 and 3 of this RFC/PATCH series are here:
https://public-inbox.org/git/20160613085546.11784-1-chriscool@xxxxxxxxxxxxx/
https://public-inbox.org/git/20160628181933.24620-1-chriscool@xxxxxxxxxxxxx/
https://public-inbox.org/git/20161130210420.15982-1-chriscool@xxxxxxxxxxxxx/
Some of the discussions related to Ben Peart's work that is used by
this series are here:
http://public-inbox.org/git/20170113155253.1644-1-benpeart@xxxxxxxxxxxxx/
http://public-inbox.org/git/20170322165220.5660-1-benpeart@xxxxxxxxxxxxx/
Links
~~~~~
This patch series is available here:
https://github.com/chriscool/git/commits/external-odb
Version 1, 2 and 3 are here:
https://github.com/chriscool/git/commits/gl-external-odb12
https://github.com/chriscool/git/commits/gl-external-odb22
https://github.com/chriscool/git/commits/gl-external-odb61
Ben Peart (4):
Documentation: add read-object-protocol.txt
contrib: add long-running-read-object/example.pl
Add t0410 to test read object mechanism
odb-helper: add read_object_process()
Christian Couder (43):
builtin/clone: get rid of 'value' strbuf
t0021/rot13-filter: refactor packet reading functions
t0021/rot13-filter: improve 'if .. elsif .. else' style
Add Git/Packet.pm from parts of t0021/rot13-filter.pl
t0021/rot13-filter: use Git/Packet.pm
Git/Packet.pm: improve error message
Git/Packet.pm: add packet_initialize()
Git/Packet: add capability functions
t0400: add 'put' command to odb-helper script
external odb: add write support
external-odb: accept only blobs for now
t0400: add test for external odb write support
Add GIT_NO_EXTERNAL_ODB env variable
Add t0410 to test external ODB transfer
lib-httpd: pass config file to start_httpd()
lib-httpd: add upload.sh
lib-httpd: add list.sh
lib-httpd: add apache-e-odb.conf
odb-helper: add 'store_plain_objects' to 'struct odb_helper'
pack-objects: don't pack objects in external odbs
t0420: add test with HTTP external odb
odb-helper: start fault in implementation
external-odb: add external_odb_fault_in_object()
odb-helper: add script_mode
external-odb: add external_odb_get_capabilities()
t04*: add 'get_cap' support to helpers
odb-helper: call odb_helper_lookup() with 'have' capability
odb-helper: fix odb_helper_fetch_object() for read_object
Add t0460 to test passing git objects
odb-helper: add read_packetized_git_object_to_fd()
odb-helper: add read_packetized_plain_object_to_fd()
Add t0470 to test passing plain objects
odb-helper: add write_object_process()
Add t0480 to test "have" capability and plain objects
external-odb: add external_odb_do_fetch_object()
odb-helper: advertise 'have' capability
odb-helper: advertise 'put' capability
odb-helper: add have_object_process()
clone: add initial param to write_remote_refs()
clone: add --initial-refspec option
clone: disable external odb before initial clone
Add test for 'clone --initial-refspec'
t: add t0430 to test cloning using bundles
Jeff King (2):
Add initial external odb support
external odb foreach
Documentation/technical/read-object-protocol.txt | 102 +++
Makefile | 2 +
builtin/clone.c | 91 ++-
builtin/pack-objects.c | 4 +
cache.h | 18 +
contrib/long-running-read-object/example.pl | 114 +++
environment.c | 4 +
external-odb.c | 220 +++++
external-odb.h | 17 +
odb-helper.c | 987 +++++++++++++++++++++++
odb-helper.h | 47 ++
perl/Git/Packet.pm | 118 +++
sha1_file.c | 117 ++-
t/lib-httpd.sh | 8 +-
t/lib-httpd/apache-e-odb.conf | 214 +++++
t/lib-httpd/list.sh | 41 +
t/lib-httpd/upload.sh | 45 ++
t/t0021/rot13-filter.pl | 97 +--
t/t0400-external-odb.sh | 85 ++
t/t0410-transfer-e-odb.sh | 148 ++++
t/t0420-transfer-http-e-odb.sh | 159 ++++
t/t0430-clone-bundle-e-odb.sh | 91 +++
t/t0450-read-object.sh | 30 +
t/t0450/read-object | 56 ++
t/t0460-read-object-git.sh | 29 +
t/t0460/read-object-git | 67 ++
t/t0470-read-object-http-e-odb.sh | 123 +++
t/t0470/read-object-plain | 93 +++
t/t0480-read-object-have-http-e-odb.sh | 123 +++
t/t0480/read-object-plain-have | 116 +++
t/t5616-clone-initial-refspec.sh | 48 ++
31 files changed, 3296 insertions(+), 118 deletions(-)
create mode 100644 Documentation/technical/read-object-protocol.txt
create mode 100644 contrib/long-running-read-object/example.pl
create mode 100644 external-odb.c
create mode 100644 external-odb.h
create mode 100644 odb-helper.c
create mode 100644 odb-helper.h
create mode 100644 perl/Git/Packet.pm
create mode 100644 t/lib-httpd/apache-e-odb.conf
create mode 100644 t/lib-httpd/list.sh
create mode 100644 t/lib-httpd/upload.sh
create mode 100755 t/t0400-external-odb.sh
create mode 100755 t/t0410-transfer-e-odb.sh
create mode 100755 t/t0420-transfer-http-e-odb.sh
create mode 100755 t/t0430-clone-bundle-e-odb.sh
create mode 100755 t/t0450-read-object.sh
create mode 100755 t/t0450/read-object
create mode 100755 t/t0460-read-object-git.sh
create mode 100755 t/t0460/read-object-git
create mode 100755 t/t0470-read-object-http-e-odb.sh
create mode 100755 t/t0470/read-object-plain
create mode 100755 t/t0480-read-object-have-http-e-odb.sh
create mode 100755 t/t0480/read-object-plain-have
create mode 100755 t/t5616-clone-initial-refspec.sh
