2019-11-13 08:37:17 +02:00

1331 lines
40 KiB
C++

/*
* Copyright (c) 2018 MariaDB Corporation Ab
*
* Use of this software is governed by the Business Source License included
* in the LICENSE.TXT file and at www.mariadb.com/bsl11.
*
* Change Date: 2023-11-12
*
* On the date above, in accordance with the Business Source License, use
* of this software will be governed by version 2 or later of the General
* Public License.
*/
#include "clustrixmonitor.hh"
#include <algorithm>
#include <set>
#include <maxbase/string.hh>
#include <maxscale/json_api.hh>
#include <maxscale/paths.h>
#include <maxscale/secrets.h>
#include <maxscale/sqlite3.h>
#include "../../../core/internal/config_runtime.hh"
#include "../../../core/internal/service.hh"
namespace http = mxb::http;
using namespace std;
using maxscale::MonitorServer;
#define LOG_JSON_ERROR(ppJson, format, ...) \
do { \
MXS_ERROR(format, ##__VA_ARGS__); \
if (ppJson) \
{ \
*ppJson = mxs_json_error_append(*ppJson, format, ##__VA_ARGS__); \
} \
} while (false)
namespace
{
namespace clustrixmon
{
config::Specification specification(MXS_MODULE_NAME, config::Specification::MONITOR);
config::ParamDuration<std::chrono::milliseconds>
cluster_monitor_interval(&specification,
"cluster_monitor_interval",
"How frequently the Clustrix monitor should perform a cluster check.",
mxs::config::INTERPRET_AS_MILLISECONDS,
std::chrono::milliseconds(DEFAULT_CLUSTER_MONITOR_INTERVAL));
config::ParamCount
health_check_threshold(&specification,
"health_check_threshold",
"How many failed health port pings before node is assumed to be down.",
DEFAULT_HEALTH_CHECK_THRESHOLD,
1, std::numeric_limits<uint32_t>::max()); // min, max
config::ParamBool
dynamic_node_detection(&specification,
"dynamic_node_detection",
"Should cluster configuration be figured out at runtime.",
DEFAULT_DYNAMIC_NODE_DETECTION);
config::ParamInteger
health_check_port(&specification,
"health_check_port",
"Port number for Clustrix health check.",
DEFAULT_HEALTH_CHECK_PORT,
0, std::numeric_limits<uint16_t>::max()); // min, max
}
const int DEFAULT_MYSQL_PORT = 3306;
const int DEFAULT_HEALTH_PORT = 3581;
// Change this, if the schema is changed.
const int SCHEMA_VERSION = 1;
static const char SQL_BN_CREATE[] =
"CREATE TABLE IF NOT EXISTS bootstrap_nodes "
"(ip CARCHAR(255), mysql_port INT)";
static const char SQL_BN_INSERT_FORMAT[] =
"INSERT INTO bootstrap_nodes (ip, mysql_port) "
"VALUES %s";
static const char SQL_BN_DELETE[] =
"DELETE FROM bootstrap_nodes";
static const char SQL_BN_SELECT[] =
"SELECT ip, mysql_port FROM bootstrap_nodes";
static const char SQL_DN_CREATE[] =
"CREATE TABLE IF NOT EXISTS dynamic_nodes "
"(id INT PRIMARY KEY, ip VARCHAR(255), mysql_port INT, health_port INT)";
static const char SQL_DN_UPSERT_FORMAT[] =
"INSERT OR REPLACE INTO dynamic_nodes (id, ip, mysql_port, health_port) "
"VALUES (%d, '%s', %d, %d)";
static const char SQL_DN_DELETE_FORMAT[] =
"DELETE FROM dynamic_nodes WHERE id = %d";
static const char SQL_DN_DELETE[] =
"DELETE FROM dynamic_nodes";
static const char SQL_DN_SELECT[] =
"SELECT ip, mysql_port FROM dynamic_nodes";
using HostPortPair = std::pair<std::string, int>;
using HostPortPairs = std::vector<HostPortPair>;
// sqlite3 callback.
int select_cb(void* pData, int nColumns, char** ppColumn, char** ppNames)
{
std::vector<HostPortPair>* pNodes = static_cast<std::vector<HostPortPair>*>(pData);
mxb_assert(nColumns == 2);
std::string host(ppColumn[0]);
int port = atoi(ppColumn[1]);
pNodes->emplace_back(host, port);
return 0;
}
}
namespace
{
bool create_schema(sqlite3* pDb)
{
char* pError = nullptr;
int rv = sqlite3_exec(pDb, SQL_BN_CREATE, nullptr, nullptr, &pError);
if (rv == SQLITE_OK)
{
rv = sqlite3_exec(pDb, SQL_DN_CREATE, nullptr, nullptr, &pError);
}
if (rv != SQLITE_OK)
{
MXS_ERROR("Could not initialize sqlite3 database: %s", pError ? pError : "Unknown error");
}
return rv == SQLITE_OK;
}
sqlite3* open_or_create_db(const std::string& path)
{
sqlite3* pDb = nullptr;
int flags = SQLITE_OPEN_READWRITE | SQLITE_OPEN_NOMUTEX | SQLITE_OPEN_CREATE;
int rv = sqlite3_open_v2(path.c_str(), &pDb, flags, nullptr);
if (rv == SQLITE_OK)
{
if (create_schema(pDb))
{
MXS_NOTICE("sqlite3 database %s open/created and initialized.", path.c_str());
}
else
{
MXS_ERROR("Could not create schema in sqlite3 database %s.", path.c_str());
if (unlink(path.c_str()) != 0)
{
MXS_ERROR("Failed to delete database %s that could not be properly "
"initialized. It should be deleted manually.", path.c_str());
sqlite3_close_v2(pDb);
pDb = nullptr;
}
}
}
else
{
if (pDb)
{
// Memory allocation failure is explained by the caller. Don't close the handle, as the
// caller will still use it even if open failed!!
MXS_ERROR("Opening/creating the sqlite3 database %s failed: %s",
path.c_str(), sqlite3_errmsg(pDb));
}
MXS_ERROR("Could not open sqlite3 database for storing information "
"about dynamically detected Clustrix nodes. The Clustrix "
"monitor will remain dependent upon statically defined "
"bootstrap nodes.");
}
return pDb;
}
}
ClustrixMonitor::Config::Config(const std::string& name)
: m_configuration(name, &clustrixmon::specification)
, m_cluster_monitor_interval(&m_configuration, &clustrixmon::cluster_monitor_interval)
, m_health_check_threshold(&m_configuration, &clustrixmon::health_check_threshold)
, m_dynamic_node_detection(&m_configuration, &clustrixmon::dynamic_node_detection)
, m_health_check_port(&m_configuration, &clustrixmon::health_check_port)
{
}
//static
void ClustrixMonitor::Config::populate(MXS_MODULE& module)
{
clustrixmon::specification.populate(module);
}
bool ClustrixMonitor::Config::configure(const MXS_CONFIG_PARAMETER& params)
{
return clustrixmon::specification.configure(m_configuration, params);
}
ClustrixMonitor::ClustrixMonitor(const string& name, const string& module, sqlite3* pDb)
: MonitorWorker(name, module)
, m_config(name)
, m_pDb(pDb)
{
}
ClustrixMonitor::~ClustrixMonitor()
{
sqlite3_close_v2(m_pDb);
}
// static
ClustrixMonitor* ClustrixMonitor::create(const string& name, const string& module)
{
string path = get_datadir();
path += "/";
path += name;
if (!mxs_mkdir_all(path.c_str(), 0744))
{
MXS_ERROR("Could not create the directory %s, MaxScale will not be "
"able to create database for persisting connection "
"information of dynamically detected Clustrix nodes.",
path.c_str());
}
path += "/clustrix_nodes-v";
path += std::to_string(SCHEMA_VERSION);
path += ".db";
sqlite3* pDb = open_or_create_db(path);
ClustrixMonitor* pThis = nullptr;
if (pDb)
{
// Even if the creation/opening of the sqlite3 database fails, we will still
// get a valid database handle.
pThis = new ClustrixMonitor(name, module, pDb);
}
else
{
// The handle will be null, *only* if the opening fails due to a memory
// allocation error.
MXS_ALERT("sqlite3 memory allocation failed, the Clustrix monitor "
"cannot continue.");
}
return pThis;
}
using std::chrono::milliseconds;
bool ClustrixMonitor::configure(const MXS_CONFIG_PARAMETER* pParams)
{
if (!clustrixmon::specification.validate(*pParams))
{
return false;
}
if (!MonitorWorker::configure(pParams))
{
return false;
}
check_bootstrap_servers();
m_health_urls.clear();
m_nodes_by_id.clear();
// Since they were validated above, failure should not be an option now.
MXB_AT_DEBUG(bool configured=) m_config.configure(*pParams);
mxb_assert(configured);
return true;
}
void ClustrixMonitor::populate_services()
{
mxb_assert(!is_running());
// The servers that the Clustrix monitor has been configured with are
// only used for bootstrapping and services will not be populated
// with them.
}
bool ClustrixMonitor::softfail(SERVER* pServer, json_t** ppError)
{
bool rv = false;
if (is_running())
{
call([this, pServer, ppError, &rv]() {
rv = perform_softfail(pServer, ppError);
},
EXECUTE_QUEUED);
}
else
{
LOG_JSON_ERROR(ppError,
"%s: The monitor is not running and hence "
"SOFTFAIL cannot be performed for %s.",
name(), pServer->address);
}
return true;
}
bool ClustrixMonitor::unsoftfail(SERVER* pServer, json_t** ppError)
{
bool rv = false;
if (is_running())
{
call([this, pServer, ppError, &rv]() {
rv = perform_unsoftfail(pServer, ppError);
},
EXECUTE_QUEUED);
}
else
{
LOG_JSON_ERROR(ppError,
"%s: The monitor is not running and hence "
"UNSOFTFAIL cannot be performed for %s.",
name(), pServer->address);
}
return true;
}
void ClustrixMonitor::server_added(SERVER* pServer)
{
// The servers explicitly added to the Cluster monitor are only used
// as bootstrap servers, so they are not added to any services.
}
void ClustrixMonitor::server_removed(SERVER* pServer)
{
// @see server_added(), no action is needed.
}
void ClustrixMonitor::pre_loop()
{
load_server_journal(nullptr);
if (m_config.dynamic_node_detection())
{
// At startup we accept softfailed nodes in an attempt to be able to
// connect at any cost. It'll be replaced once there is an alternative.
check_cluster(Clustrix::Softfailed::ACCEPT);
}
else
{
populate_from_bootstrap_servers();
}
make_health_check();
}
void ClustrixMonitor::post_loop()
{
if (m_pHub_con)
{
mysql_close(m_pHub_con);
}
m_pHub_con = nullptr;
m_pHub_server = nullptr;
}
void ClustrixMonitor::tick()
{
check_maintenance_requests();
if (m_config.dynamic_node_detection() && should_check_cluster())
{
check_cluster(Clustrix::Softfailed::REJECT);
}
switch (m_http.status())
{
case http::Async::PENDING:
MXS_WARNING("%s: Health check round had not completed when next tick arrived.", name());
break;
case http::Async::ERROR:
MXS_WARNING("%s: Health check round ended with general error.", name());
make_health_check();
break;
case http::Async::READY:
update_server_statuses();
make_health_check();
break;
}
flush_server_status();
process_state_changes();
hangup_failed_servers();
store_server_journal(nullptr);
}
void ClustrixMonitor::choose_hub(Clustrix::Softfailed softfailed)
{
mxb_assert(!m_pHub_con);
set<string> ips;
// First we check the dynamic servers, in case there are,
if (!choose_dynamic_hub(softfailed, ips))
{
// then we check the bootstrap servers, and
if (!choose_bootstrap_hub(softfailed, ips))
{
// finally, if all else fails - in practise we will only get here at
// startup (no dynamic servers) if the bootstrap servers cannot be
// contacted - we try to refresh the nodes using persisted information
if (refresh_using_persisted_nodes(ips))
{
// and then select a hub from the dynamic ones.
choose_dynamic_hub(softfailed, ips);
}
}
}
if (m_pHub_con)
{
MXS_NOTICE("%s: Monitoring Clustrix cluster state using node %s:%d.",
name(), m_pHub_server->address, m_pHub_server->port);
}
else
{
MXS_ERROR("%s: Could not connect to any server or no server that could "
"be connected to was part of the quorum.", name());
}
}
bool ClustrixMonitor::choose_dynamic_hub(Clustrix::Softfailed softfailed, std::set<string>& ips_checked)
{
for (auto& kv : m_nodes_by_id)
{
ClustrixNode& node = kv.second;
if (node.can_be_used_as_hub(name(), settings().conn_settings, softfailed))
{
m_pHub_con = node.release_connection();
m_pHub_server = node.server();
}
ips_checked.insert(node.ip());
if (m_pHub_con)
{
break;
}
}
return m_pHub_con != nullptr;
}
bool ClustrixMonitor::choose_bootstrap_hub(Clustrix::Softfailed softfailed, std::set<string>& ips_checked)
{
for (auto* pMs : servers())
{
if (ips_checked.find(pMs->server->address) == ips_checked.end())
{
if (Clustrix::ping_or_connect_to_hub(name(), settings().conn_settings, softfailed, *pMs))
{
m_pHub_con = pMs->con;
m_pHub_server = pMs->server;
}
else if (pMs->con)
{
mysql_close(pMs->con);
}
pMs->con = nullptr;
}
if (m_pHub_con)
{
break;
}
}
return m_pHub_con != nullptr;
}
bool ClustrixMonitor::refresh_using_persisted_nodes(std::set<string>& ips_checked)
{
MXS_NOTICE("Attempting to find a Clustrix bootstrap node from one of the nodes "
"used during the previous run of MaxScale.");
bool refreshed = false;
HostPortPairs nodes;
char* pError = nullptr;
int rv = sqlite3_exec(m_pDb, SQL_DN_SELECT, select_cb, &nodes, &pError);
if (rv == SQLITE_OK)
{
const std::string& username = settings().conn_settings.username;
const std::string& password = settings().conn_settings.password;
char* zPassword = decrypt_password(password.c_str());
auto it = nodes.begin();
while (!refreshed && (it != nodes.end()))
{
const auto& node = *it;
const std::string& host = node.first;
if (ips_checked.find(host) == ips_checked.end())
{
ips_checked.insert(host);
int port = node.second;
MXS_NOTICE("Trying to find out cluster nodes from %s:%d.", host.c_str(), port);
MYSQL* pHub_con = mysql_init(NULL);
if (mysql_real_connect(pHub_con, host.c_str(),
username.c_str(), zPassword,
nullptr,
port, nullptr, 0))
{
if (Clustrix::is_part_of_the_quorum(name(), pHub_con))
{
if (refresh_nodes(pHub_con))
{
MXS_NOTICE("Cluster nodes refreshed.");
refreshed = true;
}
}
else
{
MXS_WARNING("%s:%d is not part of the quorum, ignoring.", host.c_str(), port);
}
}
else
{
MXS_WARNING("Could not connect to %s:%d.", host.c_str(), port);
}
mysql_close(pHub_con);
}
++it;
}
MXS_FREE(zPassword);
}
else
{
MXS_ERROR("Could not look up persisted nodes: %s", pError ? pError : "Unknown error");
}
return refreshed;
}
bool ClustrixMonitor::refresh_nodes()
{
mxb_assert(m_pHub_con);
return refresh_nodes(m_pHub_con);
}
bool ClustrixMonitor::refresh_nodes(MYSQL* pHub_con)
{
mxb_assert(pHub_con);
map<int, ClustrixMembership> memberships;
bool refreshed = check_cluster_membership(pHub_con, &memberships);
if (refreshed)
{
const char ZQUERY[] =
"SELECT ni.nodeid, ni.iface_ip, ni.mysql_port, ni.healthmon_port, sn.nodeid "
"FROM system.nodeinfo AS ni "
"LEFT JOIN system.softfailed_nodes AS sn ON ni.nodeid = sn.nodeid";
if (mysql_query(pHub_con, ZQUERY) == 0)
{
MYSQL_RES* pResult = mysql_store_result(pHub_con);
if (pResult)
{
mxb_assert(mysql_field_count(pHub_con) == 5);
set<int> nids;
for (const auto& kv : m_nodes_by_id)
{
const ClustrixNode& node = kv.second;
nids.insert(node.id());
}
MYSQL_ROW row;
while ((row = mysql_fetch_row(pResult)) != nullptr)
{
if (row[0] && row[1])
{
int id = atoi(row[0]);
string ip = row[1];
int mysql_port = row[2] ? atoi(row[2]) : DEFAULT_MYSQL_PORT;
int health_port = row[3] ? atoi(row[3]) : DEFAULT_HEALTH_PORT;
bool softfailed = row[4] ? true : false;
// '@@' ensures no clash with user created servers.
// Monitor name ensures no clash with other Clustrix monitor instances.
string server_name = string("@@") + m_name + ":node-" + std::to_string(id);
auto nit = m_nodes_by_id.find(id);
auto mit = memberships.find(id);
if (nit != m_nodes_by_id.end())
{
// Existing node.
mxb_assert(SERVER::find_by_unique_name(server_name));
ClustrixNode& node = nit->second;
node.update(ip, mysql_port, health_port);
bool is_draining = node.server()->is_draining();
if (softfailed && !is_draining)
{
MXS_NOTICE("%s: Node %d (%s) has been SOFTFAILed. "
"Turning ON 'Being Drained'.",
name(), node.id(), node.server()->address);
node.server()->set_status(SERVER_DRAINING);
}
else if (!softfailed && is_draining)
{
MXS_NOTICE("%s: Node %d (%s) is no longer being SOFTFAILed. "
"Turning OFF 'Being Drained'.",
name(), node.id(), node.server()->address);
node.server()->clear_status(SERVER_DRAINING);
}
nids.erase(id);
}
else if (mit != memberships.end())
{
// New node.
mxb_assert(!SERVER::find_by_unique_name(server_name));
if (runtime_create_server(server_name.c_str(),
ip.c_str(),
std::to_string(mysql_port).c_str(),
"mariadbbackend",
"mysqlbackendauth",
false))
{
SERVER* pServer = SERVER::find_by_unique_name(server_name);
mxb_assert(pServer);
if (pServer)
{
if (softfailed)
{
pServer->set_status(SERVER_DRAINING);
}
const ClustrixMembership& membership = mit->second;
int health_check_threshold = m_config.health_check_threshold();
ClustrixNode node(this, membership, ip, mysql_port, health_port,
health_check_threshold, pServer);
m_nodes_by_id.insert(make_pair(id, node));
// New server, so it needs to be added to all services that
// use this monitor for defining its cluster of servers.
service_add_server(this, pServer);
}
else
{
MXS_ERROR("%s: Created server %s (at %s:%d) could not be "
"looked up using its name.",
name(), server_name.c_str(), ip.c_str(), mysql_port);
}
}
else
{
MXS_ERROR("%s: Could not create server %s at %s:%d.",
name(), server_name.c_str(), ip.c_str(), mysql_port);
}
memberships.erase(mit);
}
else
{
// Node found in system.node_info but not in system.membership
MXS_ERROR("%s: Node %d at %s:%d,%d found in system.node_info "
"but not in system.membership.",
name(), id, ip.c_str(), mysql_port, health_port);
}
}
else
{
MXS_WARNING("%s: Either nodeid and/or iface_ip is missing, ignoring node.",
name());
}
}
mysql_free_result(pResult);
// Any nodes that were not found are not available, so their
// state must be set accordingly.
for (const auto nid : nids)
{
auto it = m_nodes_by_id.find(nid);
mxb_assert(it != m_nodes_by_id.end());
ClustrixNode& node = it->second;
node.set_running(false, ClustrixNode::APPROACH_OVERRIDE);
}
cluster_checked();
}
else
{
MXS_WARNING("%s: No result returned for '%s' on %s.",
name(), ZQUERY, mysql_get_host_info(pHub_con));
}
}
else
{
MXS_ERROR("%s: Could not execute '%s' on %s: %s",
name(), ZQUERY, mysql_get_host_info(pHub_con), mysql_error(pHub_con));
}
// Since we are here, the call above to check_cluster_membership() succeeded. As that
// function may change the content of m_nodes_by_ids, we must always update the urls,
// irrespective of whether the SQL of this function succeeds or not.
update_http_urls();
}
return refreshed;
}
void ClustrixMonitor::check_bootstrap_servers()
{
HostPortPairs nodes;
char* pError = nullptr;
int rv = sqlite3_exec(m_pDb, SQL_BN_SELECT, select_cb, &nodes, &pError);
if (rv == SQLITE_OK)
{
set<string> prev_bootstrap_servers;
for (const auto& node : nodes)
{
string s = node.first + ":" + std::to_string(node.second);
prev_bootstrap_servers.insert(s);
}
set<string> current_bootstrap_servers;
for (const auto* pMs : servers())
{
SERVER* pServer = pMs->server;
string s = string(pServer->address) + ":" + std::to_string(pServer->port);
current_bootstrap_servers.insert(s);
}
if (prev_bootstrap_servers == current_bootstrap_servers)
{
MXS_NOTICE("Current bootstrap servers are the same as the ones used on "
"previous run, using persisted connection information.");
}
else if (!prev_bootstrap_servers.empty())
{
MXS_NOTICE("Current bootstrap servers (%s) are different than the ones "
"used on the previous run (%s), NOT using persistent connection "
"information.",
mxb::join(current_bootstrap_servers, ", ").c_str(),
mxb::join(prev_bootstrap_servers, ", ").c_str());
if (remove_persisted_information())
{
persist_bootstrap_servers();
}
}
}
else
{
MXS_WARNING("Could not lookup earlier bootstrap servers: %s", pError ? pError : "Unknown error");
}
}
bool ClustrixMonitor::remove_persisted_information()
{
char* pError = nullptr;
int rv;
int rv1 = sqlite3_exec(m_pDb, SQL_BN_DELETE, nullptr, nullptr, &pError);
if (rv1 != SQLITE_OK)
{
MXS_ERROR("Could not delete persisted bootstrap nodes: %s", pError ? pError : "Unknown error");
}
int rv2 = sqlite3_exec(m_pDb, SQL_DN_DELETE, nullptr, nullptr, &pError);
if (rv2 != SQLITE_OK)
{
MXS_ERROR("Could not delete persisted dynamic nodes: %s", pError ? pError : "Unknown error");
}
return rv1 == SQLITE_OK && rv2 == SQLITE_OK;
}
void ClustrixMonitor::persist_bootstrap_servers()
{
string values;
for (const auto* pMs : servers())
{
if (!values.empty())
{
values += ", ";
}
SERVER* pServer = pMs->server;
string value;
value += string("'") + pServer->address + string("'");
value += ", ";
value += std::to_string(pServer->port);
values += "(";
values += value;
values += ")";
}
if (!values.empty())
{
char insert[sizeof(SQL_BN_INSERT_FORMAT) + values.length()];
sprintf(insert, SQL_BN_INSERT_FORMAT, values.c_str());
char* pError = nullptr;
int rv = sqlite3_exec(m_pDb, insert, nullptr, nullptr, &pError);
if (rv != SQLITE_OK)
{
MXS_ERROR("Could not persist information about current bootstrap nodes: %s",
pError ? pError : "Unknown error");
}
}
}
void ClustrixMonitor::check_cluster(Clustrix::Softfailed softfailed)
{
if (m_pHub_con)
{
check_hub(softfailed);
}
if (!m_pHub_con)
{
choose_hub(softfailed);
}
if (m_pHub_con)
{
refresh_nodes();
}
}
void ClustrixMonitor::check_hub(Clustrix::Softfailed softfailed)
{
mxb_assert(m_pHub_con);
mxb_assert(m_pHub_server);
if (!Clustrix::ping_or_connect_to_hub(name(), settings().conn_settings, softfailed,
*m_pHub_server, &m_pHub_con))
{
mysql_close(m_pHub_con);
m_pHub_con = nullptr;
}
}
bool ClustrixMonitor::check_cluster_membership(MYSQL* pHub_con,
std::map<int, ClustrixMembership>* pMemberships)
{
mxb_assert(pHub_con);
mxb_assert(pMemberships);
bool rv = false;
const char ZQUERY[] = "SELECT nid, status, instance, substate FROM system.membership";
if (mysql_query(pHub_con, ZQUERY) == 0)
{
MYSQL_RES* pResult = mysql_store_result(pHub_con);
if (pResult)
{
mxb_assert(mysql_field_count(pHub_con) == 4);
set<int> nids;
for (const auto& kv : m_nodes_by_id)
{
const ClustrixNode& node = kv.second;
nids.insert(node.id());
}
MYSQL_ROW row;
while ((row = mysql_fetch_row(pResult)) != nullptr)
{
if (row[0])
{
int nid = atoi(row[0]);
string status = row[1] ? row[1] : "unknown";
int instance = row[2] ? atoi(row[2]) : -1;
string substate = row[3] ? row[3] : "unknown";
auto it = m_nodes_by_id.find(nid);
if (it != m_nodes_by_id.end())
{
ClustrixNode& node = it->second;
node.update(Clustrix::status_from_string(status),
Clustrix::substate_from_string(substate),
instance);
nids.erase(node.id());
}
else
{
ClustrixMembership membership(nid,
Clustrix::status_from_string(status),
Clustrix::substate_from_string(substate),
instance);
pMemberships->insert(make_pair(nid, membership));
}
}
else
{
MXS_WARNING("%s: No node id returned in row for '%s'.",
name(), ZQUERY);
}
}
mysql_free_result(pResult);
// Deactivate all servers that are no longer members.
for (const auto nid : nids)
{
auto it = m_nodes_by_id.find(nid);
mxb_assert(it != m_nodes_by_id.end());
ClustrixNode& node = it->second;
node.deactivate_server();
m_nodes_by_id.erase(it);
}
rv = true;
}
else
{
MXS_WARNING("%s: No result returned for '%s'.", name(), ZQUERY);
}
}
else
{
MXS_ERROR("%s: Could not execute '%s' on %s: %s",
name(), ZQUERY, mysql_get_host_info(pHub_con), mysql_error(pHub_con));
}
return rv;
}
void ClustrixMonitor::populate_from_bootstrap_servers()
{
int id = 1;
for (auto ms : servers())
{
SERVER* pServer = ms->server;
Clustrix::Status status = Clustrix::Status::UNKNOWN;
Clustrix::SubState substate = Clustrix::SubState::UNKNOWN;
int instance = 1;
ClustrixMembership membership(id, status, substate, instance);
std::string ip = pServer->address;
int mysql_port = pServer->port;
int health_port = m_config.health_check_port();
int health_check_threshold = m_config.health_check_threshold();
ClustrixNode node(this, membership, ip, mysql_port, health_port, health_check_threshold, pServer);
m_nodes_by_id.insert(make_pair(id, node));
++id;
// New server, so it needs to be added to all services that
// use this monitor for defining its cluster of servers.
service_add_server(this, pServer);
}
update_http_urls();
}
void ClustrixMonitor::update_server_statuses()
{
mxb_assert(!servers().empty());
for (auto* pMs : servers())
{
pMs->stash_current_status();
auto it = find_if(m_nodes_by_id.begin(), m_nodes_by_id.end(),
[pMs](const std::pair<int, ClustrixNode>& element) -> bool {
const ClustrixNode& info = element.second;
return pMs->server->address == info.ip();
});
if (it != m_nodes_by_id.end())
{
const ClustrixNode& info = it->second;
if (info.is_running())
{
pMs->set_pending_status(SERVER_MASTER | SERVER_RUNNING);
}
else
{
pMs->clear_pending_status(SERVER_MASTER | SERVER_RUNNING);
}
}
else
{
pMs->clear_pending_status(SERVER_MASTER | SERVER_RUNNING);
}
}
}
void ClustrixMonitor::make_health_check()
{
mxb_assert(m_http.status() != http::Async::PENDING);
m_http = mxb::http::get_async(m_health_urls);
switch (m_http.status())
{
case http::Async::PENDING:
initiate_delayed_http_check();
break;
case http::Async::ERROR:
MXS_ERROR("%s: Could not initiate health check.", name());
break;
case http::Async::READY:
MXS_INFO("%s: Health check available immediately.", name());
break;
}
}
void ClustrixMonitor::initiate_delayed_http_check()
{
mxb_assert(m_delayed_http_check_id == 0);
long max_delay_ms = settings().interval / 10;
long ms = m_http.wait_no_more_than();
if (ms > max_delay_ms)
{
ms = max_delay_ms;
}
m_delayed_http_check_id = delayed_call(ms, &ClustrixMonitor::check_http, this);
}
bool ClustrixMonitor::check_http(Call::action_t action)
{
m_delayed_http_check_id = 0;
if (action == Call::EXECUTE)
{
switch (m_http.perform())
{
case http::Async::PENDING:
initiate_delayed_http_check();
break;
case http::Async::READY:
{
mxb_assert(m_health_urls == m_http.urls());
// There are as many results as there are nodes,
// and the results are in node order.
const vector<http::Result>& results = m_http.results();
mxb_assert(results.size() == m_nodes_by_id.size());
auto it = m_nodes_by_id.begin();
for (const auto& result : results)
{
bool running = (result.code == 200); // HTTP OK
ClustrixNode& node = it->second;
node.set_running(running);
if (!running)
{
// We have to explicitly check whether the node is to be
// considered down, as the value of `health_check_threshold`
// defines how quickly a node should be considered down.
if (!node.is_running())
{
// Ok, the node is down. Trigger a cluster check at next tick.
trigger_cluster_check();
}
}
++it;
}
}
break;
case http::Async::ERROR:
MXS_ERROR("%s: Health check waiting ended with general error.", name());
}
}
return false;
}
void ClustrixMonitor::update_http_urls()
{
vector<string> health_urls;
for (const auto& kv : m_nodes_by_id)
{
const ClustrixNode& node = kv.second;
string url = "http://" + node.ip() + ":" + std::to_string(node.health_port());
health_urls.push_back(url);
}
if (m_health_urls != health_urls)
{
if (m_delayed_http_check_id != 0)
{
cancel_delayed_call(m_delayed_http_check_id);
m_delayed_http_check_id = 0;
}
m_http.reset();
m_health_urls.swap(health_urls);
}
}
bool ClustrixMonitor::perform_softfail(SERVER* pServer, json_t** ppError)
{
bool rv = perform_operation(Operation::SOFTFAIL, pServer, ppError);
// Irrespective of whether the operation succeeded or not
// a cluster check is triggered at next tick.
trigger_cluster_check();
return rv;
}
bool ClustrixMonitor::perform_unsoftfail(SERVER* pServer, json_t** ppError)
{
return perform_operation(Operation::UNSOFTFAIL, pServer, ppError);
}
bool ClustrixMonitor::perform_operation(Operation operation,
SERVER* pServer,
json_t** ppError)
{
bool performed = false;
const char ZSOFTFAIL[] = "SOFTFAIL";
const char ZUNSOFTFAIL[] = "UNSOFTFAIL";
const char* zOperation = (operation == Operation::SOFTFAIL) ? ZSOFTFAIL : ZUNSOFTFAIL;
if (!m_pHub_con)
{
check_cluster(Clustrix::Softfailed::ACCEPT);
}
if (m_pHub_con)
{
auto it = find_if(m_nodes_by_id.begin(), m_nodes_by_id.end(),
[pServer](const std::pair<int, ClustrixNode>& element) {
return element.second.server() == pServer;
});
if (it != m_nodes_by_id.end())
{
ClustrixNode& node = it->second;
const char ZQUERY_FORMAT[] = "ALTER CLUSTER %s %d";
int id = node.id();
// ZUNSOFTFAIL is longer
char zQuery[sizeof(ZQUERY_FORMAT) + sizeof(ZUNSOFTFAIL) + UINTLEN(id)];
sprintf(zQuery, ZQUERY_FORMAT, zOperation, id);
if (mysql_query(m_pHub_con, zQuery) == 0)
{
MXS_NOTICE("%s: %s performed on node %d (%s).",
name(), zOperation, id, pServer->address);
if (operation == Operation::SOFTFAIL)
{
MXS_NOTICE("%s: Turning on 'Being Drained' on server %s.",
name(), pServer->address);
pServer->set_status(SERVER_DRAINING);
}
else
{
mxb_assert(operation == Operation::UNSOFTFAIL);
MXS_NOTICE("%s: Turning off 'Being Drained' on server %s.",
name(), pServer->address);
pServer->clear_status(SERVER_DRAINING);
}
}
else
{
LOG_JSON_ERROR(ppError,
"%s: The execution of '%s' failed: %s",
name(), zQuery, mysql_error(m_pHub_con));
}
}
else
{
LOG_JSON_ERROR(ppError,
"%s: The server %s is not being monitored, "
"cannot perform %s.",
name(), pServer->address, zOperation);
}
}
else
{
LOG_JSON_ERROR(ppError,
"%s: Could not could not connect to any Clustrix node, "
"cannot perform %s of %s.",
name(), zOperation, pServer->address);
}
return performed;
}
void ClustrixMonitor::persist(const ClustrixNode& node)
{
if (!m_pDb)
{
return;
}
char sql_upsert[sizeof(SQL_DN_UPSERT_FORMAT) + 10 + node.ip().length() + 10 + 10];
int id = node.id();
const char* zIp = node.ip().c_str();
int mysql_port = node.mysql_port();
int health_port = node.health_port();
sprintf(sql_upsert, SQL_DN_UPSERT_FORMAT, id, zIp, mysql_port, health_port);
char* pError = nullptr;
if (sqlite3_exec(m_pDb, sql_upsert, nullptr, nullptr, &pError) == SQLITE_OK)
{
MXS_INFO("Updated Clustrix node in bookkeeping: %d, '%s', %d, %d.",
id, zIp, mysql_port, health_port);
}
else
{
MXS_ERROR("Could not update Ćlustrix node (%d, '%s', %d, %d) in bookkeeping: %s",
id, zIp, mysql_port, health_port, pError ? pError : "Unknown error");
}
}
void ClustrixMonitor::unpersist(const ClustrixNode& node)
{
if (!m_pDb)
{
return;
}
char sql_delete[sizeof(SQL_DN_UPSERT_FORMAT) + 10];
int id = node.id();
sprintf(sql_delete, SQL_DN_DELETE_FORMAT, id);
char* pError = nullptr;
if (sqlite3_exec(m_pDb, sql_delete, nullptr, nullptr, &pError) == SQLITE_OK)
{
MXS_INFO("Deleted Clustrix node %d from bookkeeping.", id);
}
else
{
MXS_ERROR("Could not delete Ćlustrix node %d from bookkeeping: %s",
id, pError ? pError : "Unknown error");
}
}