rgw: fix busy wait in RGWHTTPManager

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: sync local mdlog if needed

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: wake up wait coroutine thread on destruction

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

Add zone create admin command - wip

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: add id to RGWZone

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: poll mdlog for incremental sync

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: read mdlog entries for incremental sync

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: get remote mdlog status when initializing

Use that as the lowerbound for the meta marker when we start
the incremental sync.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: hash full sync keys to shards

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: switch to incremental sync when done full meta sync

each shard need to collect the coroutines it spawned before continuing.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: update global sync state after building meta indexes

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: update meta sync markers during full sync operation

update marker per shard, keep a window so that we don't update after each change

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: add zonegroup name or id to zonegroup get and set commands

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: Add zonegroup create, delete and rename admin commands

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: add period activate admin command and fix get current command

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: add current_period to RGWRealm

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: RGWZoneGroup should inheirt from RGWSystemMetaObj

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: Add old_format support to RGWSystemMetaObj

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: add id to zonegroup

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: remove region

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: change RGWZoneGroup::get_pool_name to return string

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: convert old region configuration to zonegroups during init

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

fix ceph-dencoder link

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

cmake: add rgw_sync.cc and rgw_coroutine.cc

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: fix async sleep and wakeup logic

lot's of cleanups and fixes

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: full meta sync

need to handle errors, and update markers, but full meta sync is functional

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: naively fetch all meta objects

doing it one by one, later will spawn these concurrently

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: read remote meta coroutine

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: read full sync indexes

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: modify sync state, sync run uses state

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: coroutie for omap_get_vals()

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: keep track of spawned coroutines under the current op

instead of keeping a per-stack list of spawned coroutines, keep it
on the op that spawned it. Otherwise, we might spawn a cr, call a second
one that will spawn some crs, and when waiting for these to complete
it will also wait on the first one.
The list of spawned coroutines is inherited. Once a coroutine finishes
its parent will take over all the spawned ones that weren't collected.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: replace omap_set coroutine code

No need to use the async rados worker.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: fix the simple coroutine

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: fix async sleep and wakeup logic

lot's of cleanups and fixes

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: coroutines producer consumers

Create a new producer-consumer(s) abstraction for the coroutines
infrastructure.
Move environment info from the op into the stack. Stack can be
sent to sleep and awaken explicitly.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: build index of all master zone meta keys

still wip, but getting there

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: create a coroutine for fetching REST api objects

this simplifies the code that yanks all the metadata from the master

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: fetch all metadata entries

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: meta sync cleanups

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: remove calls for synchronous rados ops

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

radosgw-admin: dump shard status on mdlog sync status

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: initialize all shard markers

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: return error if any spawned op errored

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: iterate through all mdlog sync status shards

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: remove unused code, cleanups

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: adjust mdlog fetch to use new interfaces

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: fix init mdlog sync

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: mdlog sync init

initialize mdlog sync state, uses the implicit yield scheme (from boost)

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: use boost for stackless coroutine flow

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: add cookie to lock coroutines

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: more lock/unlock coroutines work

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: prepare rados lock operation to be used with coroutines

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: simple rgw metadata write coroutine

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: move coroutines code

move the generic code to rgw_coroutine.{h,cc}

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: return retcode of failing coroutine

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: rename async ops to coroutines

which they basically are

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: fail if master zone tries to use meta sync

We'll change it later, but as it is now that's the way to handle it
gracefully.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: Fix init_complete

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: support old region configurations

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: Remove region admin commands

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: Add zone group admin commands

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: zonegroup json encoding/decoding with backward compatability

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: Rename Region to ZoneGroup only internal structures and class

User visuable are not changed: admin commands, jsons and rados objects names

Signed-off-by: Orit Wasserman <owasserm@redhat.com>
Conflicts:
src/rgw/rgw_admin.cc
src/rgw/rgw_rados.cc

rgw: Add period admin commands and rest api to handle period

the new commands are prepare,delete, get, activate, pull and push

Signed-off-by: Orit Wasserman <owasserm@redhat.com>
Conflicts:
src/rgw/rgw_admin.cc

rgw: Add RGWPeriod

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: Add RGWRealm class and admin commands to handle it

admin realm commands: create, get, delete, rename, set-default and get-default

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: Use switch for raw storage opcodes

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: use string for default realm

Conflicts:
src/rgw/rgw_rados.cc

rgw: Add RGWSystemMetaObj class

Signed-off-by: Orit Wasserman <owasserm@redhat.com>
Conflicts:
src/rgw/rgw_rados.cc

rgw: Add RGWNameToId and RGWDefaultSystemMetaObjInfo

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: fix test_rgw_admin_log.cc

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

cmake: add rgw_sync.cc

Signed-off-by: Orit Wasserman <owasserm@redhat.com>

rgw: async ops can generate new stacks for parallel execution

Can now have an op generating another op that will be called at
a different stack. The current op will halt execution until new
op finishes. We can call this multiple times so that multiple ops
will be executed concurrently. This can be useful in the case we
want to generate a single op that will read from multiple objects,
after it had read some meta information. So the root op will read
the meta information and will generate multiple request that will
execute concurrently (albeit on a single thread).

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: a few bug fixes related to async operations

fix uninitialized state, trim stacks, etc.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: add a generic op to pull data off metadata obj

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: add another state for simple async ops

so that these ops can chain other ops and handle their
response

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: more simplification

create a helper method in the ops manager to run a single op, it
builds the needed env

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: more simplification, no need to take reference to op

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: some cleanup, keep async env structure

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: create simple async op handler

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: async rados requests through a work queue

Tie in the async rados requests to the async framework.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: retrieve a specific sync shard info

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw-admin: command to show mdlog sync status

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: stackable ops infrastructure

ops can now call other ops, so that we could simplify
the different states each ops need to keep.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: initial work on sync status marker

Also, rename radosgw-admin mdlog sync to radosgw-admin mdlog-fetch.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: read mdlog marker

Use our local marker as a starting point for fetching the mater mdlog.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: clean up mdlog clone states, other fixes

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: don't reference msg after finish_request()

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: hide the completion manager

some more layer separation

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: create some abstraction around async ops infrastructure

This will be reused later for other operations.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: mdlog sync error reporting on failed request

Basic error reporting. We'll need to dump this to the backend and
add admin commands to read.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

some error handling work, not ready yet

rgw: md log cloning completely async

hook librados completions into the rgw completion notifier, split the
md log write into two different states.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: rework metadata log sync to use async capabilities

We now have a window of operations. The next step will be to
asynchronously write to the local mdlog, so that we don't
just hold the work on a synchronous request.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: can now wait on async rest requests

In addition to waiting on a specific request, added a
completion manager that we can wait on so that we get
which request is ready.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: add new class to create resource request

RGWRESTReadResource holds the request input and output. Previously
we were using stack variables that couldn't be used in the async
path.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: more http aio stuff

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: threaded http manager groundwork

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: more work on cleaning up http client interfaces

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>