[teammgrd]: Improve LAGs cleanup on shutdown: send SIGTERM directly t…

…o PID. (#1841)

This PR is intended to fix LAGs cleanup degradation caused by python2.7 -> python3 migration.
The approach is to replace `teamd -k -t` call with the raw `SIGTERM` and add PID alive check.
This will make sure the `teammgrd` is stopped only after all managed processes are being killed.

resolves: #8071

**What I did**
* Replaced `teamd -k -t` call with raw `SIGTERM`
* Added PID alive check

**Why I did it**
* To fix LAGs cleanup timeout issue caused by python2.7 -> python3 upgrade

**How I verified it**
1. Configure 64 LAG RIFs
2. Reload config

cfgmgr/teammgr.cpp

@@ -112,18 +112,53 @@ void TeamMgr::doTask(Consumer &consumer)
     }
 }
 
-
 void TeamMgr::cleanTeamProcesses()
 {
     SWSS_LOG_ENTER();
     SWSS_LOG_NOTICE("Cleaning up LAGs during shutdown...");
-    for (const auto& it: m_lagList)
+
+    std::unordered_map<std::string, pid_t> aliasPidMap;
+
+    for (const auto& alias: m_lagList)
+    {
+        std::string res;
+        pid_t pid;
+
+        {
+            std::stringstream cmd;
+            cmd << "cat " << shellquote("/var/run/teamd/" + alias + ".pid");
+            EXEC_WITH_ERROR_THROW(cmd.str(), res);
+
+            pid = static_cast<pid_t>(std::stoul(res, nullptr, 10));
+            aliasPidMap[alias] = pid;
+
+            SWSS_LOG_INFO("Read port channel %s pid %d", alias.c_str(), pid);
+        }
+
+        {
+            std::stringstream cmd;
+            cmd << "kill -TERM " << pid;
+            EXEC_WITH_ERROR_THROW(cmd.str(), res);
+
+            SWSS_LOG_INFO("Sent SIGTERM to port channel %s pid %d", alias.c_str(), pid);
+        }
+    }
+
+    for (const auto& cit: aliasPidMap)
     {
-        //This will call team -k kill -t <teamdevicename> which internally send SIGTERM 
-        removeLag(it);
+        const auto &alias = cit.first;
+        const auto &pid = cit.second;
+
+        std::stringstream cmd;
+        std::string res;
+
+        SWSS_LOG_NOTICE("Waiting for port channel %s pid %d to stop...", alias.c_str(), pid);
+
+        cmd << "tail -f --pid=" << pid << " /dev/null";
+        EXEC_WITH_ERROR_THROW(cmd.str(), res);
     }
 
-    return;
+    SWSS_LOG_NOTICE("LAGs cleanup is done");
 }
 
 void TeamMgr::doLagTask(Consumer &consumer)

cfgmgr/teammgr.h

@@ -32,7 +32,6 @@ class TeamMgr : public Orch
     ProducerStateTable m_appLagTable;
 
     std::set<std::string> m_lagList;
-    std::map<std::string, pid_t> m_lagPIDList;
 
     MacAddress m_mac;
 
@@ -50,7 +49,6 @@ class TeamMgr : public Orch
     bool setLagMtu(const std::string &alias, const std::string &mtu);
     bool setLagLearnMode(const std::string &alias, const std::string &learn_mode);
     bool setLagTpid(const std::string &alias, const std::string &tpid);
-
 
     bool isPortEnslaved(const std::string &);
     bool findPortMaster(std::string &, const std::string &);

cfgmgr/teammgrd.cpp

@@ -66,7 +66,7 @@ int main(int argc, char **argv)
         }
 
         while (!received_sigterm)
-        {            
+        {
             Selectable *sel;
             int ret;
 
@@ -91,7 +91,8 @@ int main(int argc, char **argv)
     catch (const exception &e)
     {
         SWSS_LOG_ERROR("Runtime error: %s", e.what());
+        return EXIT_FAILURE;
     }
 
-    return -1;
+    return EXIT_SUCCESS;
 }