2
0
mirror of https://github.com/openvswitch/ovs synced 2025-08-22 01:51:26 +00:00
ovs/ovsdb/server.c
Ben Pfaff 1b1d2e6daa ovsdb: Introduce experimental support for clustered databases.
This commit adds support for OVSDB clustering via Raft.  Please read
ovsdb(7) for information on how to set up a clustered database.  It is
simple and boils down to running "ovsdb-tool create-cluster" on one server
and "ovsdb-tool join-cluster" on each of the others and then starting
ovsdb-server in the usual way on all of them.

One you have a clustered database, you configure ovn-controller and
ovn-northd to use it by pointing them to all of the servers, e.g. where
previously you might have said "tcp:1.2.3.4" was the database server,
now you say that it is "tcp:1.2.3.4,tcp:5.6.7.8,tcp:9.10.11.12".

This also adds support for database clustering to ovs-sandbox.

Acked-by: Justin Pettit <jpettit@ovn.org>
Tested-by: aginwala <aginwala@asu.edu>
Signed-off-by: Ben Pfaff <blp@ovn.org>
2018-03-24 12:04:53 -07:00

219 lines
6.9 KiB
C

/* Copyright (c) 2011, 2012 Nicira, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <config.h>
#include "server.h"
#include "hash.h"
#include "ovsdb.h"
#include "uuid.h"
/* Initializes 'session' as a session within 'server'. */
void
ovsdb_session_init(struct ovsdb_session *session, struct ovsdb_server *server)
{
session->server = server;
ovs_list_init(&session->completions);
hmap_init(&session->waiters);
}
/* Destroys 'session'. */
void
ovsdb_session_destroy(struct ovsdb_session *session)
{
ovs_assert(hmap_is_empty(&session->waiters));
hmap_destroy(&session->waiters);
}
/* Searches 'session' for an ovsdb_lock_waiter named 'lock_name' and returns
* it if it finds one, otherwise NULL. */
struct ovsdb_lock_waiter *
ovsdb_session_get_lock_waiter(const struct ovsdb_session *session,
const char *lock_name)
{
struct ovsdb_lock_waiter *waiter;
HMAP_FOR_EACH_WITH_HASH (waiter, session_node, hash_string(lock_name, 0),
&session->waiters) {
if (!strcmp(lock_name, waiter->lock_name)) {
return waiter;
}
}
return NULL;
}
/* Returns the waiter that owns 'lock'.
*
* A lock always has an owner, so this function will never return NULL. */
struct ovsdb_lock_waiter *
ovsdb_lock_get_owner(const struct ovsdb_lock *lock)
{
return CONTAINER_OF(ovs_list_front(&lock->waiters),
struct ovsdb_lock_waiter, lock_node);
}
/* Removes 'waiter' from its lock's list. This means that, if 'waiter' was
* formerly the owner of its lock, then it no longer owns it.
*
* Returns the session that now owns 'waiter'. This is NULL if 'waiter' was
* the lock's owner and no other sessions were waiting for the lock. In this
* case, the lock has been destroyed, so the caller must be sure not to refer
* to it again. A nonnull return value reflects a change in the lock's
* ownership if and only if 'waiter' formerly owned the lock. */
struct ovsdb_session *
ovsdb_lock_waiter_remove(struct ovsdb_lock_waiter *waiter)
{
struct ovsdb_lock *lock = waiter->lock;
ovs_list_remove(&waiter->lock_node);
waiter->lock = NULL;
if (ovs_list_is_empty(&lock->waiters)) {
hmap_remove(&lock->server->locks, &lock->hmap_node);
free(lock->name);
free(lock);
return NULL;
}
return ovsdb_lock_get_owner(lock)->session;
}
/* Destroys 'waiter', which must have already been removed from its lock's
* waiting list with ovsdb_lock_waiter_remove().
*
* Removing and destroying locks are decoupled because a lock initially created
* by the "steal" request, that is later stolen by another client, remains in
* the database session until the database client sends an "unlock" request. */
void
ovsdb_lock_waiter_destroy(struct ovsdb_lock_waiter *waiter)
{
ovs_assert(!waiter->lock);
hmap_remove(&waiter->session->waiters, &waiter->session_node);
free(waiter->lock_name);
free(waiter);
}
/* Returns true if 'waiter' owns its associated lock. */
bool
ovsdb_lock_waiter_is_owner(const struct ovsdb_lock_waiter *waiter)
{
return waiter->lock && waiter == ovsdb_lock_get_owner(waiter->lock);
}
/* Initializes 'server'.
*
* The caller must call ovsdb_server_add_db() for each database to which
* 'server' should provide access. */
void
ovsdb_server_init(struct ovsdb_server *server)
{
shash_init(&server->dbs);
hmap_init(&server->locks);
uuid_generate(&server->uuid);
}
/* Adds 'db' to the set of databases served out by 'server'. Returns true if
* successful, false if 'db''s name is the same as some database already in
* 'server'. */
bool
ovsdb_server_add_db(struct ovsdb_server *server, struct ovsdb *db)
{
return shash_add_once(&server->dbs, db->name, db);
}
/* Removes 'db' from the set of databases served out by 'server'. */
void
ovsdb_server_remove_db(struct ovsdb_server *server, struct ovsdb *db)
{
shash_find_and_delete_assert(&server->dbs, db->name);
}
/* Destroys 'server'. */
void
ovsdb_server_destroy(struct ovsdb_server *server)
{
shash_destroy(&server->dbs);
hmap_destroy(&server->locks);
}
static struct ovsdb_lock *
ovsdb_server_create_lock__(struct ovsdb_server *server, const char *lock_name,
uint32_t hash)
{
struct ovsdb_lock *lock;
HMAP_FOR_EACH_WITH_HASH (lock, hmap_node, hash, &server->locks) {
if (!strcmp(lock->name, lock_name)) {
return lock;
}
}
lock = xzalloc(sizeof *lock);
lock->server = server;
lock->name = xstrdup(lock_name);
hmap_insert(&server->locks, &lock->hmap_node, hash);
ovs_list_init(&lock->waiters);
return lock;
}
/* Attempts to acquire the lock named 'lock_name' for 'session' within
* 'server'. Returns the new lock waiter.
*
* If 'mode' is OVSDB_LOCK_STEAL, then the new lock waiter is always the owner
* of the lock. '*victimp' receives the session of the previous owner or NULL
* if the lock was previously unowned. (If the victim itself originally
* obtained the lock through a "steal" operation, then this function also
* removes the victim from the lock's waiting list.)
*
* If 'mode' is OVSDB_LOCK_WAIT, then the new lock waiter is the owner of the
* lock only if this lock had no existing owner. '*victimp' is set to NULL. */
struct ovsdb_lock_waiter *
ovsdb_server_lock(struct ovsdb_server *server,
struct ovsdb_session *session,
const char *lock_name,
enum ovsdb_lock_mode mode,
struct ovsdb_session **victimp)
{
uint32_t hash = hash_string(lock_name, 0);
struct ovsdb_lock_waiter *waiter, *victim;
struct ovsdb_lock *lock;
lock = ovsdb_server_create_lock__(server, lock_name, hash);
victim = (mode == OVSDB_LOCK_STEAL && !ovs_list_is_empty(&lock->waiters)
? ovsdb_lock_get_owner(lock)
: NULL);
waiter = xmalloc(sizeof *waiter);
waiter->mode = mode;
waiter->lock_name = xstrdup(lock_name);
waiter->lock = lock;
if (mode == OVSDB_LOCK_STEAL) {
ovs_list_push_front(&lock->waiters, &waiter->lock_node);
} else {
ovs_list_push_back(&lock->waiters, &waiter->lock_node);
}
waiter->session = session;
hmap_insert(&waiter->session->waiters, &waiter->session_node, hash);
if (victim && victim->mode == OVSDB_LOCK_STEAL) {
ovsdb_lock_waiter_remove(victim);
}
*victimp = victim ? victim->session : NULL;
return waiter;
}