It's been confusing for a long time that EC pools are implemented by
ReplicatedPG. What PG/ReplicatedPG really implement is the concept
of a PG where consistency is managed by the primary via a log.
Signed-off-by: Samuel Just <sjust@redhat.com>
The prototype of the init functions of erasure coded plugins is changed
from
int __erasure_code_init(char *plugin_name)
to
int __erasure_code_init(char *plugin_name, char *directory)
The jerasure plugin will find optimized variants in this directory and
load them. The load() and preload() functions of
ErasureCodePluginRegistry only use a directory instead of a more generic
parameters map. The parameters map was only used for the directory entry
anyway.
Signed-off-by: Loic Dachary <loic-201408@dachary.org>
* Warn the reader that the implementation is ahead and may differ
* Update the links to the Firefly branch
* Remove links to issues used during development to avoid confusion
Signed-off-by: Loic Dachary <loic@dachary.org>
* The parameters come from the erasure code profile
* Add a note about the upstream submodules gf-complete / jerasure
Signed-off-by: Loic Dachary <loic@dachary.org>
Update the introduction to explain erasure code profiles. Remove
obsolete explanations about partial writes etc. Remove links to tickets
used during development. Update permalinks to be closer to
Firefly (v0.78).
Signed-off-by: Loic Dachary <loic@dachary.org>
Based on a dialog with Sam ( as published at http://dachary.org/?p=2320 ).
* Remove PGBackend-h.rst because PGBackend.h is now in master.
* Fix typos caught by ispell
* Update recovery links to point to PGBackend recover methods
* Workaround formating warning
developer_notes.rst:3: WARNING: Duplicate explicit target name:
"erasurecodepluginexample" which should be legitimate.
Signed-off-by: Loic Dachary <loic@dachary.org>
* Update to the current state of the ghobject implementaiton and the fact
that they encode the shard_t Although the pool also contains the shard
id, it is less relevant to understand the implementation.
* Update with the erasure code plugin infrastructure and the example
plugin now in master.
* Move jerasure to a separate page to be expanded and link it from the
toc
* Kill the partial read and writes notes as it will probably not be
implemented in the near future. Kill some of the notes because they
are no longer relevant.
* Add a definition for "chunk rank"
* Reword, update schemas, fix typos.
Signed-off-by: Loic Dachary <loic@dachary.org>
* unify conventions to match those used by jerasure ( data chunk = K,
coding chunk = M, use coding instead of parity, use erasures instead
of erased )
* make lines 80 characters long
* modify the descriptions to take into account that the chunk rank
will encoded in the pool name and not on a per object basis
* remove the doxygen link to ErasureCodeInterface because it fails
doc: asphyxiate does not support class
http://tracker.ceph.com/issues/6115
* only systematic codes are considered at this point ( all jerasure
techniques are systematic). Although the API could be extended to
include non systematic codes, it is probably a case of over
engineering at this point.
* add link to
http://tracker.ceph.com/issues/6113
add ceph osd pool create [name] [key=value]
* update the plugin system description to match the proposed
implementation http://tracker.ceph.com/issues/5877http://tracker.ceph.com/issues/4929 refs #4929
Reviewed-by: Joao Eduardo Luis <joao.luis@inktank.com>
Signed-off-by: Loic Dachary <loic@dachary.org>
* replace the erasure code plugin abstract interface with a doxygen link
that will be populated when the header shows in master
* update the plugin documentation to reflect the current draft implementation
* fix broken link to PGBackend-h
* add a glossary to define chunk, stripe, shard and strip with a drawing
http://tracker.ceph.com/issues/4929 refs #4929
Signed-off-by: Loic Dachary <loic@dachary.org>
Explains how objects are stored and used in erasure coded pools. It is
the result of discussions that occured on the ceph-devel mailing list
around june 2013. The rationale behind each change can be found in the
archive of the mailing list. For instance, the coding of the chunk
number with the object or the decision to decode using any K chunks
instead of trying to fetch the data chunks when possible because it
would allow simple concatenation when systematic codes are used.
http://tracker.ceph.com/issues/4929 refs #4929
Signed-off-by: Loic Dachary <loic@dachary.org>
