hcq/MaxScale
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

MXS-2220 Move server response calculation functions inside class

Browse Source
This commit is contained in:
Esa Korhonen
2019-01-08 10:17:07 +02:00
parent df17619722
commit 9cac927542
5 changed files with 45 additions and 73 deletions
Download Patch File Download Diff File Expand all files Collapse all files

40
include/maxscale/server.hh
View File

@ -14,9 +14,11 @@
#include <maxscale/ccdefs.hh>
#include <mutex>
#include <string>
#include <maxscale/ssl.h>
#include <unordered_map>
#include <maxbase/average.hh>
#include <maxscale/ssl.h>
// A mapping from a path to a percentage, e.g.: "/disk" -> 80.
typedef std::unordered_map<std::string, int32_t> MxsDiskSpaceThreshold;
@ -474,27 +476,35 @@ public:
*/
void clear_status(uint64_t bit);
int response_time_num_samples() const
{
return m_response_time.num_samples();
}
double response_time_average() const
{
return m_response_time.average();
}
/**
* Add a response time measurement to the global server value.
*
* @param ave The value to add
* @param num_samples The weight of the new value, that is, the number of measurement points it represents
*/
void response_time_add(double ave, int num_samples);
protected:
SERVER()
: m_response_time(maxbase::EMAverage {0.04, 0.35, 500})
{
}
private:
static const int DEFAULT_CHARSET = 0x08; /** The latin1 charset */
static const int DEFAULT_CHARSET = 0x08; /**< The latin1 charset */
maxbase::EMAverage m_response_time; /**< Response time calculations for this server */
std::mutex m_average_write_mutex; /**< Protects response time from concurrent writing */
};
/**
* @brief Add a response average to the server response average.
*
* @param server The server.
* @param ave Average.
* @param num_samples Number of samples the average consists of.
*
*/
void server_add_response_average(SERVER* server, double ave, int num_samples);
int server_response_time_num_samples(const SERVER* server);
double server_response_time_average(const SERVER* server);
namespace maxscale
{

15
server/core/internal/server.hh
View File

@ -20,7 +20,6 @@
#include <map>
#include <mutex>
#include <maxbase/average.hh>
#include <maxscale/config.hh>
#include <maxscale/server.hh>
#include <maxscale/resultset.hh>
@ -34,7 +33,6 @@ public:
Server(const std::string& name, const std::string& protocol = "", const std::string& authenticator = "")
: SERVER()
, m_name(name)
, m_response_time(maxbase::EMAverage {0.04, 0.35, 500})
{
m_settings.protocol = protocol;
m_settings.authenticator = authenticator;
@ -46,18 +44,6 @@ public:
std::string value;
};
int response_time_num_samples() const
{
return m_response_time.num_samples();
}
double response_time_average() const
{
return m_response_time.average();
}
void response_time_add(double ave, int num_samples);
long persistpoolmax() const
{
return m_settings.persistpoolmax;
@ -371,7 +357,6 @@ private:
const std::string m_name; /**< Server config name */
Settings m_settings; /**< Server settings */
VersionInfo info; /**< Server version and type information */
maxbase::EMAverage m_response_time; /**< Response time calculations for this server */
};
/**

49
server/core/server.cc
View File

@ -74,7 +74,6 @@ namespace
struct ThisUnit
{
std::mutex ave_write_mutex; /**< TODO: Move to Server */
std::mutex all_servers_lock; /**< Protects access to all_servers */
std::list<Server*> all_servers; /**< Global list of all servers */
} this_unit;
@ -517,9 +516,9 @@ void Server::print_to_dcb(DCB* dcb) const
dcb_printf(dcb, "\tCurrent no. of operations: %d\n", server->stats.n_current_ops);
dcb_printf(dcb, "\tNumber of routed packets: %lu\n", server->stats.packets);
std::ostringstream ave_os;
if (server_response_time_num_samples(server))
if (response_time_num_samples())
{
maxbase::Duration dur(server_response_time_average(server));
maxbase::Duration dur(response_time_average());
ave_os << dur;
}
else
@ -1101,7 +1100,7 @@ json_t* Server::server_json_attributes(const Server* server)
json_object_set_new(stats, "active_operations", json_integer(server->stats.n_current_ops));
json_object_set_new(stats, "routed_packets", json_integer(server->stats.packets));
maxbase::Duration response_ave(server_response_time_average(server));
maxbase::Duration response_ave(server->response_time_average());
json_object_set_new(stats, "adaptive_avg_select_time", json_string(to_string(response_ave).c_str()));
json_object_set_new(attr, "statistics", stats);
@ -1174,38 +1173,20 @@ bool Server::set_disk_space_threshold(const char* disk_space_threshold)
return rv;
}
void server_add_response_average(SERVER* srv, double ave, int num_samples)
{
Server* server = static_cast<Server*>(srv);
Guard guard(this_unit.ave_write_mutex);
server->response_time_add(ave, num_samples);
}
int server_response_time_num_samples(const SERVER* srv)
{
const Server* server = static_cast<const Server*>(srv);
return server->response_time_num_samples();
}
double server_response_time_average(const SERVER* srv)
{
const Server* server = static_cast<const Server*>(srv);
return server->response_time_average();
}
/** Apply backend average and adjust sample_max, which determines the weight of a new average
* applied to EMAverage.
* Sample max is raised if the server is fast, aggresively lowered if the incoming average is clearly
* lower than the EMA, else just lowered a bit. The normal increase and decrease, drifting, of the max
* is done to follow the speed of a server. The important part is the lowering of max, to allow for a
* server that is speeding up to be adjusted and used.
*
* Three new magic numbers to replace the sample max magic number...
*
*/
void Server::response_time_add(double ave, int num_samples)
void SERVER::response_time_add(double ave, int num_samples)
{
/**
* Apply backend average and adjust sample_max, which determines the weight of a new average
* applied to EMAverage.
*
* Sample max is raised if the server is fast, aggressively lowered if the incoming average is clearly
* lower than the EMA, else just lowered a bit. The normal increase and decrease, drifting, of the max
* is done to follow the speed of a server. The important part is the lowering of max, to allow for a
* server that is speeding up to be adjusted and used.
*
* Three new magic numbers to replace the sample max magic number... */
constexpr double drift {1.1};
Guard guard(m_average_write_mutex);
int current_max = m_response_time.sample_max();
int new_max {0};

4
server/modules/routing/readwritesplit/rwsplit_select_backends.cc
View File

@ -129,7 +129,7 @@ PRWBackends::iterator backend_cmp_response_time(PRWBackends& sBackends)
for (int i = 0; i < SZ; ++i)
{
SERVER_REF* server = (*sBackends[i]).backend();
double ave = server_response_time_average(server->server);
double ave = server->server->response_time_average();
if (ave == 0)
{
@ -302,7 +302,7 @@ static void log_server_connections(select_criteria_t criteria, const PRWBackends
case ADAPTIVE_ROUTING:
{
maxbase::Duration response_ave(server_response_time_average(b->server));
maxbase::Duration response_ave(b->server->response_time_average());
std::ostringstream os;
os << response_ave;
MXS_INFO("adaptive avg. select time: %s from \t[%s]:%d %s",

10
server/modules/routing/readwritesplit/rwsplitsession.cc
View File

@ -120,9 +120,7 @@ void RWSplitSession::close()
if (stat.make_valid())
{
server_add_response_average(backend->server(),
stat.average().secs(),
stat.num_samples());
backend->server()->response_time_add(stat.average().secs(), stat.num_samples());
}
backend->response_stat().reset();
@ -637,11 +635,9 @@ void RWSplitSession::clientReply(GWBUF* writebuf, DCB* backend_dcb)
ResponseStat& stat = backend->response_stat();
stat.query_ended();
if (stat.is_valid() && (stat.sync_time_reached()
|| server_response_time_num_samples(backend->server()) == 0))
|| backend->server()->response_time_num_samples() == 0))
{
server_add_response_average(backend->server(),
stat.average().secs(),
stat.num_samples());
backend->server()->response_time_add(stat.average().secs(), stat.num_samples());
stat.reset();
}