Fix bug: some trial jobs hyper params not stored and error handling updates (#225)

* Pull latest code (#2)

* webui logpath and document (#135)

* Add webui document and logpath as a href

* fix tslint

* fix comments by Chengmin

* Pai training service bug fix and enhancement (#136)

* Add NNI installation scripts

* Update pai script, update NNI_out_dir

* Update NNI dir in nni sdk local.py

* Create .nni folder in nni sdk local.py

* Add check before creating .nni folder

* Fix typo for PAI_INSTALL_NNI_SHELL_FORMAT

* Improve annotation (#138)

* Improve annotation

* Minor bugfix

* Selectively install through pip (#139)

Selectively install through pip 
* update setup.py

* fix paiTrainingService bugs (#137)

* fix nnictl bug

* add hdfs host validation

* fix bugs

* fix dockerfile

* fix install.sh

* update install.sh

* fix dockerfile

* Set timeout for HDFSUtility exists function

* remove unused TODO

* fix sdk

* add optional for outputDir and dataDir

* refactor dockerfile.base

* Remove unused import in hdfsclientUtility

* Add documentation for NNI PAI mode experiment (#141)

* Add documentation for NNI PAI mode

* Fix typo based on PR comments

* Exit with subprocess return code of trial keeper

* Remove additional exit code

* Fix typo based on PR comments

* update doc for smac tuner (#140)

* Revert "Selectively install through pip (#139)" due to potential pip install issue (#142)

* Revert "Selectively install through pip (#139)"

This reverts commit 1d174836.

* Add exit code of subprocess for trial_keeper

* Update README, add link to PAImode doc

* fix bug (#147)

* Refactor nnictl and add config_pai.yml (#144)

* fix nnictl bug

* add hdfs host validation

* fix bugs

* fix dockerfile

* fix install.sh

* update install.sh

* fix dockerfile

* Set timeout for HDFSUtility exists function

* remove unused TODO

* fix sdk

* add optional for outputDir and dataDir

* refactor dockerfile.base

* Remove unused import in hdfsclientUtility

* add config_pai.yml

* refactor nnictl create logic and add colorful print

* fix nnictl stop logic

* add annotation for config_pai.yml

* add document for start experiment

* fix config.yml

* fix document

* Fix trial keeper wrongly exit issue (#152)

* Fix trial keeper bug, use actual exitcode to exit rather than 1

* Fix bug of table sort (#145)

* Update doc for PAIMode and v0.2 release notes (#153)

* Update v0.2 documentation regards to release note and PAI training service

* Update document to describe NNI docker image

* Bug fix for SQuAD example tuner. (#134)

* Update Makefile (#151)

* test

* update setup.py

* update Makefile and install.sh

* rever setup.py

* change color

* update doc

* update doc

* fix auto-completion's extra space

* update Makefile

* update webui

* Update doc image (#163)

* update doc

* trivial

* trivial

* trivial

* trivial

* trivial

* trivial

* update image

* update image size

* Update ga squad (#104)

* update readme in ga_squad

* update readme

* fix typo

* Update README.md

* Update README.md

* Update README.md

* update readme

* sklearn examples (#169)

* fix nnictl bug

* fix install.sh

* add sklearn-regression example

* add sklearn classification

* update sklearn

* update example

* remove additional code

* Update batch tuner (#158)

* update readme in ga_squad

* update readme

* fix typo

* Update README.md

* Update README.md

* Update README.md

* update readme

* update batch tuner

* Quickly fix cascading search space bug in tuner (#156)

* update readme in ga_squad

* update readme

* fix typo

* Update README.md

* Update README.md

* Update README.md

* update readme

* quickly fix cascading searchspace bug in tuner

* Add iterative search space example (#119)

* update readme in ga_squad

* update readme

* fix typo

* Update README.md

* Update README.md

* Update README.md

* update readme

* add iterative search space example

* update

* update readme

* change name

* Fix bug: some trial jobs hyper params not stored when fast finished

* updates

* updates

* updates

* updates

src/nni_manager/common/errors.ts

@@ -26,9 +26,12 @@ export namespace NNIErrorNames {
 }
 
 export class NNIError extends Error {
-    constructor (name: string, message: string) {
+    constructor (name: string, message: string, err?: Error) {
         super(message);
         this.name = name;
+        if (err !== undefined) {
+            this.stack = err.stack;
+        }
     }
 }
 

src/nni_manager/core/ipcInterface.ts

@@ -23,6 +23,7 @@ import * as assert from 'assert';
 import { ChildProcess } from 'child_process';
 import { EventEmitter } from 'events';
 import { Readable, Writable } from 'stream';
+import { NNIError } from '../common/errors';
 import { getLogger, Logger } from '../common/log';
 import * as CommandType from './commands';
 
@@ -98,10 +99,15 @@ class IpcInterface {
     public sendCommand(commandType: string, content: string = ''): void {
         this.logger.debug(`ipcInterface command type: [${commandType}], content:[${content}]`);
         assert.ok(this.acceptCommandTypes.has(commandType));
+
+        try {
             const data: Buffer = encodeCommand(commandType, content);
             if (!this.outgoingStream.write(data)) {
                 this.logger.error('Commands jammed in buffer!');
             }
+        } catch (err) {
+            throw new NNIError('Dispatcher Error', `Dispatcher Error: ${err.message}`, err);
+        }
     }
 
     /**

src/nni_manager/core/nniDataStore.ts

@@ -25,8 +25,8 @@ import { Deferred } from 'ts-deferred';
 import * as component from '../common/component';
 import { Database, DataStore, MetricData, MetricDataRecord, MetricType,
     TrialJobEvent, TrialJobEventRecord, TrialJobInfo } from '../common/datastore';
-import { isNewExperiment } from '../common/experimentStartupInfo';
-import { getExperimentId } from '../common/experimentStartupInfo';
+import { NNIError } from '../common/errors';
+import { getExperimentId, isNewExperiment } from '../common/experimentStartupInfo';
 import { getLogger, Logger } from '../common/log';
 import { ExperimentProfile,  TrialJobStatistics } from '../common/manager';
 import { TrialJobStatus } from '../common/trainingService';
@@ -72,7 +72,11 @@ class NNIDataStore implements DataStore {
     }
 
     public async storeExperimentProfile(experimentProfile: ExperimentProfile): Promise<void> {
+        try {
             await this.db.storeExperimentProfile(experimentProfile);
+        } catch (err) {
+            throw new NNIError('Datastore error', `Datastore error: ${err.message}`, err);
+        }
     }
 
     public getExperimentProfile(experimentId: string): Promise<ExperimentProfile> {
@@ -82,7 +86,11 @@ class NNIDataStore implements DataStore {
     public storeTrialJobEvent(event: TrialJobEvent, trialJobId: string, data?: string, logPath?: string): Promise<void> {
         this.log.debug(`storeTrialJobEvent: event: ${event}, data: ${data}, logpath: ${logPath}`);
 
-        return this.db.storeTrialJobEvent(event, trialJobId, data, logPath);
+        return this.db.storeTrialJobEvent(event, trialJobId, data, logPath).catch(
+                (err: Error) => {
+                    throw new NNIError('Datastore error', `Datastore error: ${err.message}`, err);
+                }
+            );
     }
 
     public async getTrialJobStatistics(): Promise<any[]> {
@@ -128,6 +136,7 @@ class NNIDataStore implements DataStore {
             return;
         }
         assert(trialJobId === metrics.trial_job_id);
+        try {
             await this.db.storeMetricData(trialJobId, JSON.stringify({
                 trialJobId: metrics.trial_job_id,
                 parameterId: metrics.parameter_id,
@@ -136,6 +145,9 @@ class NNIDataStore implements DataStore {
                 data: metrics.value,
                 timestamp: Date.now()
             }));
+        } catch (err) {
+            throw new NNIError('Datastore error', `Datastore error: ${err.message}`, err);
+        }
     }
 
     public getMetricData(trialJobId?: string, metricType?: MetricType): Promise<MetricDataRecord[]> {

src/nni_manager/core/nnimanager.ts

@@ -25,6 +25,7 @@ import { ChildProcess, spawn } from 'child_process';
 import { Deferred } from 'ts-deferred';
 import * as component from '../common/component';
 import { DataStore, MetricDataRecord, MetricType, TrialJobInfo } from '../common/datastore';
+import { NNIError } from '../common/errors';
 import { getExperimentId } from '../common/experimentStartupInfo';
 import { getLogger, Logger } from '../common/log';
 import {
@@ -122,7 +123,7 @@ class NNIManager implements Manager {
         }
 
         const dispatcherCommand: string = getMsgDispatcherCommand(expParams.tuner, expParams.assessor, expParams.multiPhase);
-        console.log(`dispatcher command: ${dispatcherCommand}`);
+        this.log.debug(`dispatcher command: ${dispatcherCommand}`);
         this.setupTuner(
             //expParams.tuner.tunerCommand,
             dispatcherCommand,
@@ -136,6 +137,7 @@ class NNIManager implements Manager {
         this.run().catch((err: Error) => {
             this.criticalError(err);
         });
+
         return this.experimentProfile.id;
     }
 
@@ -146,12 +148,12 @@ class NNIManager implements Manager {
         const expParams: ExperimentParams = this.experimentProfile.params;
 
         // Set up multiphase config
-        if(expParams.multiPhase && this.trainingService.isMultiPhaseJobSupported) {
+        if (expParams.multiPhase && this.trainingService.isMultiPhaseJobSupported) {
             this.trainingService.setClusterMetadata('multiPhase', expParams.multiPhase.toString());
         }
 
         const dispatcherCommand: string = getMsgDispatcherCommand(expParams.tuner, expParams.assessor, expParams.multiPhase);
-        console.log(`dispatcher command: ${dispatcherCommand}`);
+        this.log.debug(`dispatcher command: ${dispatcherCommand}`);
         this.setupTuner(
             dispatcherCommand,
             undefined,
@@ -369,8 +371,12 @@ class NNIManager implements Manager {
 
         await Promise.all([
             this.periodicallyUpdateExecDuration(),
-            this.trainingService.run(),
-            this.trialJobsMaintainer.run()]);
+            this.trainingService.run().catch((err: Error) => {
+                throw new NNIError('Training service error', `Training service error: ${err.message}`, err);
+            }),
+            this.trialJobsMaintainer.run().catch((err: Error) => {
+                throw new NNIError('Job maintainer error', `Job maintainer error: ${err.message}`, err);
+            })]);
     }
 
      private addEventListeners(): void {
@@ -380,26 +386,26 @@ class NNIManager implements Manager {
         }
         this.trainingService.addTrialJobMetricListener((metric: TrialJobMetric) => {
             this.onTrialJobMetrics(metric).catch((err: Error) => {
-                this.criticalError(err);
+                this.criticalError(new NNIError('Job metrics error', `Job metrics error: ${err.message}`, err));
             });
         });
 
         this.trialJobsMaintainer.on(async (event: TrialJobMaintainerEvent, trialJobDetail: TrialJobDetail) => {
             this.onTrialJobEvent(event, trialJobDetail).catch((err: Error) => {
-                this.criticalError(err);
+                this.criticalError(new NNIError('Trial job event error', `Trial job event error: ${err.message}`, err));
             });
         });
 
         this.dispatcher.onCommand((commandType: string, content: string) => {
             this.onTunerCommand(commandType, content).catch((err: Error) => {
-                this.criticalError(err);
+                this.criticalError(new NNIError('Tuner command event error', `Tuner command event error: ${err.message}`, err));
             });
         });
     }
 
     private sendInitTunerCommands(): void {
         if (this.dispatcher === undefined) {
-            throw new Error('Error: tuner has not been setup');
+            throw new Error('Dispatcher error: tuner has not been setup');
         }
         // TO DO: we should send INITIALIZE command to tuner if user's tuner needs to run init method in tuner
         this.log.debug(`Send tuner command: update search space: ${this.experimentProfile.params.searchSpace}`);
@@ -479,9 +485,13 @@ class NNIManager implements Manager {
                     };
                     const trialJobDetail: TrialJobDetail = await this.trainingService.submitTrialJob(trialJobAppForm);
                     this.trialJobsMaintainer.setTrialJob(trialJobDetail.id, Object.assign({}, trialJobDetail));
-                    // TO DO: to uncomment
-                    assert(trialJobDetail.status === 'WAITING');
-                    await this.dataStore.storeTrialJobEvent(trialJobDetail.status, trialJobDetail.id, content, trialJobDetail.url);
+                    const jobDetailSnapshot: TrialJobDetail | undefined = this.trialJobsMaintainer.getTrialJob(trialJobDetail.id);
+                    if (jobDetailSnapshot !== undefined) {
+                        await this.dataStore.storeTrialJobEvent(
+                            jobDetailSnapshot.status, jobDetailSnapshot.id, content, jobDetailSnapshot.url);
+                    } else {
+                        assert(false, `undefined jobdetail in job maintainer: ${trialJobDetail.id}`);
+                    }
                     if (this.currSubmittedTrialNum === this.experimentProfile.params.maxTrialNum) {
                         this.trialJobsMaintainer.setNoMoreTrials();
                     }