From 8c93890bd39ed05d7109619da57dc4675b0fd28a Mon Sep 17 00:00:00 2001
From: QuanluZhang <z.quanluzhang@gmail.com>
Date: Fri, 26 Oct 2018 09:28:40 +0800
Subject: [PATCH] fix bug about execDuration and endTime (#270)

* fix bug about execDuration and endTime

* modify time interval to 30 seconds

* refactor based on Gems's suggestion

* for triggering ci
---
 src/nni_manager/core/nnimanager.ts | 31 +++++++++++++++---------------
 1 file changed, 15 insertions(+), 16 deletions(-)

diff --git a/src/nni_manager/core/nnimanager.ts b/src/nni_manager/core/nnimanager.ts
index 19f5afe08e..796236c670 100644
--- a/src/nni_manager/core/nnimanager.ts
+++ b/src/nni_manager/core/nnimanager.ts
@@ -48,7 +48,7 @@ import { createDispatcherInterface, IpcInterface } from './ipcInterface';
 class NNIManager implements Manager {
     private trainingService: TrainingService;
     private dispatcher: IpcInterface | undefined;
-    private currSubmittedTrialNum: number; // need to be recovered
+    private currSubmittedTrialNum: number;  // need to be recovered
     private trialConcurrencyChange: number; // >0: increase, <0: decrease
     private customizedTrials: string[]; // need to be recovered
     private log: Logger;
@@ -58,7 +58,6 @@ class NNIManager implements Manager {
     private status: NNIManagerStatus;
     private waitingTrials: string[];
     private trialJobs: Map<string, TrialJobDetail>;
-    private suspendDuration: number;
 
     constructor() {
         this.currSubmittedTrialNum = 0;
@@ -69,7 +68,6 @@ class NNIManager implements Manager {
         this.dispatcherPid = 0;
         this.waitingTrials = [];
         this.trialJobs = new Map<string, TrialJobDetail>();
-        this.suspendDuration = 0;
 
         this.log = getLogger();
         this.dataStore = component.get(DataStore);
@@ -336,12 +334,16 @@ class NNIManager implements Manager {
     }
 
     private async periodicallyUpdateExecDuration(): Promise<void> {
-        const startTime: number = Date.now();
-        const execDuration: number = this.experimentProfile.execDuration;
+        let count: number = 1;
         for (; ;) {
-            await delay(1000 * 60 * 10); // 10 minutes
-            this.experimentProfile.execDuration = execDuration + (Date.now() - startTime) / 1000 - this.suspendDuration;
-            await this.storeExperimentProfile();
+            await delay(1000 * 1); // 1 seconds
+            if (this.status.status === 'EXPERIMENT_RUNNING') {
+                this.experimentProfile.execDuration += 1;
+                if (count % 10 === 0) {
+                    await this.storeExperimentProfile();
+                }
+            }
+            count += 1;
         }
     }
 
@@ -351,7 +353,6 @@ class NNIManager implements Manager {
         for (const trialJobId of Array.from(this.trialJobs.keys())) {
             const trialJobDetail: TrialJobDetail = await this.trainingService.getTrialJob(trialJobId);
             const oldTrialJobDetail: TrialJobDetail | undefined = this.trialJobs.get(trialJobId);
-            //assert(oldTrialJobDetail);
             if (oldTrialJobDetail !== undefined && oldTrialJobDetail.status !== trialJobDetail.status) {
                 this.trialJobs.set(trialJobId, Object.assign({}, trialJobDetail));
                 await this.dataStore.storeTrialJobEvent(trialJobDetail.status, trialJobDetail.id, undefined, trialJobDetail.url);
@@ -388,8 +389,6 @@ class NNIManager implements Manager {
             throw new Error('Error: tuner has not been setup');
         }
         let allFinishedTrialJobNum: number = 0;
-        const startTime: number = Date.now();
-        let suspendStartTime: number = 0;
         for (; ;) {
             if (this.status.status === 'STOPPING') {
                 break;
@@ -426,18 +425,18 @@ class NNIManager implements Manager {
             }
 
             // check maxtrialnum and maxduration here
-            if ((Date.now() - startTime) / 1000 + this.experimentProfile.execDuration - this.suspendDuration 
-                > this.experimentProfile.params.maxExecDuration ||
+            if (this.experimentProfile.execDuration > this.experimentProfile.params.maxExecDuration ||
                 this.currSubmittedTrialNum >= this.experimentProfile.params.maxTrialNum) {
                 assert(this.status.status === 'EXPERIMENT_RUNNING' || this.status.status === 'DONE');
                 if (this.status.status === 'EXPERIMENT_RUNNING') {
-                    suspendStartTime = Date.now();
+                    this.experimentProfile.endTime = Date.now();
+                    await this.storeExperimentProfile();
                 }
                 this.status.status = 'DONE';
             } else {
                 if (this.status.status === 'DONE') {
-                    assert(suspendStartTime !== 0);
-                    this.suspendDuration += (Date.now() - suspendStartTime) / 1000;
+                    delete this.experimentProfile.endTime;
+                    await this.storeExperimentProfile();
                 }
                 this.status.status = 'EXPERIMENT_RUNNING';
                 for (let i: number = this.trialJobs.size; i < this.experimentProfile.params.trialConcurrency; i++) {