1
0
mirror of https://github.com/mariadb-corporation/mariadb-columnstore-engine.git synced 2025-07-30 19:23:07 +03:00

MCOL-1523

This commit is contained in:
David Hill
2018-09-05 14:53:13 -05:00
parent e5f18964f0
commit 8b0507b987
5 changed files with 52 additions and 106 deletions

View File

@ -7036,17 +7036,8 @@ int processCommand(string* arguments)
if (systemstatus.SystemOpState == oam::ACTIVE ) { if (systemstatus.SystemOpState == oam::ACTIVE ) {
try try
{ {
// cout << endl << " Starting Modules" << endl;
// oam.startModule(devicenetworklist, ackTemp);
//reload DBRM with new configuration, needs to be done here after startModule
// cmd = startup::StartUp::installDir() + "/bin/dbrmctl reload > /dev/null 2>&1";
// system(cmd.c_str());
// sleep(15);
// cout << " Successful start of Modules " << endl;
cout << endl << " Restarting System "; cout << endl << " Restarting System ";
gracefulTemp = oam::FORCEFUL;
int returnStatus = oam.restartSystem(gracefulTemp, ackTemp); int returnStatus = oam.restartSystem(gracefulTemp, ackTemp);
switch (returnStatus) switch (returnStatus)
{ {

View File

@ -3172,7 +3172,7 @@ int main(int argc, char *argv[])
//skip interface with remote servers and perform install //skip interface with remote servers and perform install
if ( !nonDistribute ) if ( !nonDistribute )
{ {
// //
// perform remote install of other servers in the system // perform remote install of other servers in the system
// //
cout << endl << "===== System Installation =====" << endl << endl; cout << endl << "===== System Installation =====" << endl << endl;

View File

@ -1489,7 +1489,7 @@ void pingDeviceThread()
if (moduleInfoList[moduleName] >= ModuleHeartbeatCount || if (moduleInfoList[moduleName] >= ModuleHeartbeatCount ||
opState == oam::DOWN || opState == oam::AUTO_DISABLED) opState == oam::DOWN || opState == oam::AUTO_DISABLED)
{ {
log.writeLog(__LINE__, "Module alive, bring it back online: " + moduleName, LOG_TYPE_DEBUG); log.writeLog(__LINE__, "*** Module alive, bring it back online: " + moduleName, LOG_TYPE_DEBUG);
string PrimaryUMModuleName = config.moduleName(); string PrimaryUMModuleName = config.moduleName();
try { try {
@ -1927,7 +1927,7 @@ void pingDeviceThread()
{ {
//Log failure, issue alarm, set moduleOpState //Log failure, issue alarm, set moduleOpState
Configuration config; Configuration config;
log.writeLog(__LINE__, "module is down: " + moduleName, LOG_TYPE_CRITICAL); log.writeLog(__LINE__, "*** module is down: " + moduleName, LOG_TYPE_CRITICAL);
//set query system state not ready //set query system state not ready
BRM::DBRM dbrm; BRM::DBRM dbrm;
@ -2013,9 +2013,6 @@ void pingDeviceThread()
// resume the dbrm // resume the dbrm
oam.dbrmctl("resume"); oam.dbrmctl("resume");
log.writeLog(__LINE__, "'dbrmctl resume' done", LOG_TYPE_DEBUG); log.writeLog(__LINE__, "'dbrmctl resume' done", LOG_TYPE_DEBUG);
//set recycle process
processManager.recycleProcess(moduleName);
} }
// return values = 'ip address' for running or rebooting, stopped or terminated // return values = 'ip address' for running or rebooting, stopped or terminated
@ -2234,9 +2231,6 @@ void pingDeviceThread()
oam.dbrmctl("resume"); oam.dbrmctl("resume");
log.writeLog(__LINE__, "'dbrmctl resume' done", LOG_TYPE_DEBUG); log.writeLog(__LINE__, "'dbrmctl resume' done", LOG_TYPE_DEBUG);
//set recycle process
processManager.recycleProcess(moduleName);
//enable query stats //enable query stats
dbrm.setSystemQueryReady(true); dbrm.setSystemQueryReady(true);

View File

@ -619,21 +619,11 @@ void processMSG(messageqcpp::IOSocket* cfIos)
//distribute config file //distribute config file
processManager.distributeConfigFile("system"); processManager.distributeConfigFile("system");
processManager.setSystemState(oam::ACTIVE);
//set query system state ready //set query system state ready
processManager.setQuerySystemState(true); processManager.setQuerySystemState(true);
//call dbrm control processManager.setSystemState(oam::ACTIVE);
/* oam.dbrmctl("halt"); }
log.writeLog(__LINE__, "'dbrmctl halt' done", LOG_TYPE_DEBUG);
oam.dbrmctl("reload");
log.writeLog(__LINE__, "'dbrmctl reload' done", LOG_TYPE_DEBUG);
oam.dbrmctl("resume");
log.writeLog(__LINE__, "'dbrmctl resume' done", LOG_TYPE_DEBUG);
*/ }
} }
else else
{ {
@ -846,15 +836,10 @@ void processMSG(messageqcpp::IOSocket* cfIos)
status = processManager.disableModule(moduleName, true); status = processManager.disableModule(moduleName, true);
log.writeLog(__LINE__, "Disable Module Completed on " + moduleName, LOG_TYPE_INFO); log.writeLog(__LINE__, "Disable Module Completed on " + moduleName, LOG_TYPE_INFO);
processManager.recycleProcess(moduleName);
//check for SIMPLEX Processes on mate might need to be started
processManager.checkSimplexModule(moduleName);
processManager.setSystemState(oam::ACTIVE);
//set query system state ready //set query system state ready
processManager.setQuerySystemState(true); processManager.setQuerySystemState(true);
processManager.setSystemState(oam::ACTIVE);
} }
else else
{ {
@ -1611,6 +1596,7 @@ void processMSG(messageqcpp::IOSocket* cfIos)
status = retStatus; status = retStatus;
} }
} }
//now stop local module //now stop local module
processManager.stopModule(config.moduleName(), graceful, manualFlag ); processManager.stopModule(config.moduleName(), graceful, manualFlag );
@ -1627,7 +1613,7 @@ void processMSG(messageqcpp::IOSocket* cfIos)
oam::DeviceNetworkList devicenetworklist; oam::DeviceNetworkList devicenetworklist;
pthread_t startsystemthread; pthread_t startsystemthread;
pthread_create (&startsystemthread, NULL, (void*(*)(void*)) &startSystemThread, &devicenetworklist); status = pthread_create (&startsystemthread, NULL, (void*(*)(void*)) &startSystemThread, &devicenetworklist);
if ( status != 0 ) { if ( status != 0 ) {
log.writeLog(__LINE__, "STARTMODULE: pthread_create failed, return status = " + oam.itoa(status)); log.writeLog(__LINE__, "STARTMODULE: pthread_create failed, return status = " + oam.itoa(status));
@ -1636,20 +1622,19 @@ void processMSG(messageqcpp::IOSocket* cfIos)
if (status == 0 && ackIndicator) if (status == 0 && ackIndicator)
{ {
// BUG 4554 We don't need the join because calpont console is now looking for "Active" pthread_join(startsystemthread, NULL);
// We need to return the ack right away to let console know we got the message. status = startsystemthreadStatus;
// pthread_join(startsystemthread, NULL);
// status = startsystemthreadStatus;
} }
// setup MySQL Replication after switchover command // setup MySQL Replication after FORCE restart command
/* if (graceful == FORCEFUL) if ( (status == API_SUCCESS) &&
(graceful == oam::FORCEFUL) )
{ {
log.writeLog(__LINE__, "Setup MySQL Replication for restartSystem FORCE, used by switch-parent command", LOG_TYPE_DEBUG); log.writeLog(__LINE__, "Setup MySQL Replication for restartSystem FORCE", LOG_TYPE_DEBUG);
oam::DeviceNetworkList devicenetworklist; oam::DeviceNetworkList devicenetworklist;
processManager.setMySQLReplication(devicenetworklist); processManager.setMySQLReplication(devicenetworklist);
} }
*/
log.writeLog(__LINE__, "RESTARTSYSTEM: Start System Request Completed", LOG_TYPE_INFO); log.writeLog(__LINE__, "RESTARTSYSTEM: Start System Request Completed", LOG_TYPE_INFO);
} }
@ -3277,6 +3262,7 @@ int ProcessManager::shutdownModule(string target, ByteStream::byte actionIndicat
int ProcessManager::disableModule(string target, bool manualFlag) int ProcessManager::disableModule(string target, bool manualFlag)
{ {
Oam oam; Oam oam;
ProcessManager processManager(config, log);
ModuleConfig moduleconfig; ModuleConfig moduleconfig;
log.writeLog(__LINE__, "disableModule request for " + target, LOG_TYPE_DEBUG); log.writeLog(__LINE__, "disableModule request for " + target, LOG_TYPE_DEBUG);
@ -3386,6 +3372,11 @@ int ProcessManager::disableModule(string target, bool manualFlag)
if ( updateWorkerNodeconfig() != API_SUCCESS ) if ( updateWorkerNodeconfig() != API_SUCCESS )
return API_FAILURE; return API_FAILURE;
processManager.recycleProcess(target);
//check for SIMPLEX Processes on mate might need to be started
processManager.checkSimplexModule(target);
//distribute config file //distribute config file
distributeConfigFile("system"); distributeConfigFile("system");
@ -3414,37 +3405,6 @@ void ProcessManager::recycleProcess(string module, bool enableModule)
oam.getSystemConfig("PrimaryUMModuleName", PrimaryUMModuleName); oam.getSystemConfig("PrimaryUMModuleName", PrimaryUMModuleName);
} }
catch(...) {} catch(...) {}
// restart DBRM Process and DMLProc and return if enable module is being done
if (enableModule)
{
//recycle DBRM processes in all cases
restartProcessType("DBRMControllerNode");
restartProcessType("DBRMWorkerNode");
sleep(5);
restartProcessType("DMLProc");
return;
}
//recycle DBRM processes in all cases
// restartProcessType("DBRMControllerNode", module);
// restartProcessType("DBRMWorkerNode");
// only recycle ddl/dmlproc, if down/up module is non-parent UM
/* if ( ( moduleType == "um" ) &&
if ( PrimaryUMModuleName != module)
{
restartProcessType("DDLProc",module);
restartProcessType("DMLProc",module);
return;
}
*/
// if( PrimaryUMModuleName == module)
// {
// stopProcessType("DDLProc");
// stopProcessType("DMLProc");
// }
stopProcessType("WriteEngineServer"); stopProcessType("WriteEngineServer");
@ -3526,10 +3486,6 @@ int ProcessManager::enableModule(string target, int state, bool failover)
if ( newStandbyModule == target) if ( newStandbyModule == target)
setStandbyModule(newStandbyModule); setStandbyModule(newStandbyModule);
//set recycle process
// if (!failover)
// recycleProcess(target);
log.writeLog(__LINE__, "enableModule request for " + target + " completed", LOG_TYPE_DEBUG); log.writeLog(__LINE__, "enableModule request for " + target + " completed", LOG_TYPE_DEBUG);
@ -6518,15 +6474,15 @@ void ProcessManager::setQuerySystemState(bool set)
Oam oam; Oam oam;
BRM::DBRM dbrm; BRM::DBRM dbrm;
log.writeLog(__LINE__, "setQuerySystemState = " + oam.itoa(set), LOG_TYPE_DEBUG); log.writeLog(__LINE__, "setQuerySystemState called = " + oam.itoa(set), LOG_TYPE_DEBUG);
try { try {
dbrm.setSystemQueryReady(set); dbrm.setSystemQueryReady(set);
log.writeLog(__LINE__, "setSystemQueryReady successful", LOG_TYPE_DEBUG); log.writeLog(__LINE__, "setSystemQueryReady = " + oam.itoa(set), LOG_TYPE_DEBUG);
try { try {
dbrm.setSystemReady(set); dbrm.setSystemReady(set);
log.writeLog(__LINE__, "setSystemReady successful", LOG_TYPE_DEBUG); log.writeLog(__LINE__, "setSystemReady = " + oam.itoa(set), LOG_TYPE_DEBUG);
} }
catch(...) catch(...)
{ {
@ -7089,23 +7045,28 @@ void startSystemThread(oam::DeviceNetworkList Devicenetworklist)
} }
if (DMLprocessstatus.ProcessOpState == oam::ACTIVE) { if (DMLprocessstatus.ProcessOpState == oam::ACTIVE) {
rtn = oam::ACTIVE; rtn = oam::ACTIVE;
break; break;
} }
if (DMLprocessstatus.ProcessOpState == oam::FAILED) { if (DMLprocessstatus.ProcessOpState == oam::FAILED) {
rtn = oam::FAILED; rtn = oam::FAILED;
status = oam::API_FAILURE;
break; break;
} }
// wait some more // wait some more
sleep(2); sleep(2);
} }
processManager.setSystemState(rtn);
if ( rtn = oam::ACTIVE )
//set query system state not ready
processManager.setQuerySystemState(true);
processManager.setSystemState(rtn);
} }
else
//set query system state ready processManager.setSystemState(oam::FAILED);
processManager.setQuerySystemState(true);
// exit thread // exit thread
log.writeLog(__LINE__, "startSystemThread Exit", LOG_TYPE_DEBUG); log.writeLog(__LINE__, "startSystemThread Exit", LOG_TYPE_DEBUG);

View File

@ -2451,17 +2451,17 @@ void processStatusMSG(messageqcpp::IOSocket* cfIos)
} }
//if DMLProc set to ACTIVE, set system state to ACTIVE if in an INIT state //if DMLProc set to ACTIVE, set system state to ACTIVE if in an INIT state
if ( processName == "DMLProc" && state == oam::ACTIVE ) // if ( processName == "DMLProc" && state == oam::ACTIVE )
{ // {
if ( fShmSystemStatus[0].OpState == oam::BUSY_INIT || // if ( fShmSystemStatus[0].OpState == oam::BUSY_INIT ||
fShmSystemStatus[0].OpState == oam::MAN_INIT || // fShmSystemStatus[0].OpState == oam::MAN_INIT ||
fShmSystemStatus[0].OpState == oam::AUTO_INIT ) // fShmSystemStatus[0].OpState == oam::AUTO_INIT )
{ // {
fShmSystemStatus[0].OpState = state; // fShmSystemStatus[0].OpState = state;
memcpy(fShmSystemStatus[0].StateChangeDate, oam.getCurrentTime().c_str(), DATESIZE); // memcpy(fShmSystemStatus[0].StateChangeDate, oam.getCurrentTime().c_str(), DATESIZE);
log.writeLog(__LINE__, "statusControl: REQUEST RECEIVED: Set System State = " + oamState[state], LOG_TYPE_DEBUG); // log.writeLog(__LINE__, "statusControl: REQUEST RECEIVED: Set System State = " + oamState[state], LOG_TYPE_DEBUG);
} // }
} // }
} }
break; break;