Problem:
I am working on an Android app which interacts with nodejs REST server using node orm for mysql backend. On my server, I have a functionality of authenticating users based on email verification. Once verification is successful, node orm fetches the user object, changes the verified column value and saves it back.
But, the change is not reflecting in the db after execution. Only if we run the same code another time, it is reflecting in the database
Code
exports.activateEmail = function(email, callback) {
log.info('In verifyEmailDao.js, activateEmail module');
var db = connectionObj.getConnection();
var Candidate = db.models.jobseeker_id;
Candidate.find({email : email}, function(err,candidate){
if(err){
log.info('cannot find account with email to activate', email);
callback(false, null);
}
else {
candidate[0].verified = true;
log.info('candidate email now activated.! status is', candidate[0].verified);
candidate[0].save(function(error){
log.info('Email verified any errors?', error);
callback(true, candidate[0].id);
});
}
});
}
Edit 1:
jobseeker_id.js (node-orm model)
var orm = require('orm');
module.exports = function(db){
console.log('coming inside candidateId.js');
var JobSeekerId = db.define('jobseeker_id', {
id : {type:'serial' , key:true},
first_name : String,
last_name : String,
email : String,
password : String,
verified : Boolean
},{
validations : {
email : orm.enforce.unique("Already registered")
}
});
}
Server log:
{"name":"test-app" "msg":"In verifyEmailDao.js, activateEmail module"}
{"name":"test-app","msg":"candidate email now activated.! status is true"}
{"name":"test-app","msg":"Email verified any errors? null"}
{"name":"test-app","msg":"Email sucessfully activated. Now deleting the entry from verify email link table for candidate id 30}
{"name":"test-app","msg":"In verifyEmailDao.js, deleteRandomLink module"}
{"name":"test-app","msg":"error is---> null"}
{"name":"test-app","msg":"Entry deleted from verify email table as email is activated"}
There will no be no changes in the log when I execute the code for second time, but the change in the db will be reflected!
After 2 days of hell I finally fixed the issue by adding a statement db.settings.set('instance.cache', false) to the db config file. Though I did'nt clearly understand how db update issue was resolved by setting the cache to false, this did the trick!
Related
I am having the following problem with node and mysql:
I have a function registerUser that takes the req.body with the user credentials and store them into a mysql db.
First of all I check that the email provided does not already exist. I have done this validation working with postgres in the following manner: if(user.rows.lenght!==0) return res.send("user already exist")
Then I pass to the next line of code that insrts the credentials into the db.
My problem is that using mysql, user.rows is undefined. I am having trouble extracting the data from the response which would allowme to perform some sort of validation.
My code is like this:
registerUser:async(req,res)=>{
const resolver=Resolver(res)
try {
//get data from req.body
const {userName, userEmail, userPassword}=req.body
//Check if user alreday exist on db by email
const user=db.query("SELECT * FROM users WHERE user_email=?",
[userEmail],(err,result)=>{
if(err) console.log(err)
else if(result.length!==0) return res.status(401).send('user already exist')
})
The callback function of the query does not stop the execution of the registeruser function. Also, the result comes with the user credentaials which is what I need, but I dont know how to extract it from the callback in order to use it in the scope of registerUser
I'm trying to implement a feature that detects whether or not an account has already signed up. Now when doing parallel requests I'm getting a deadlock. I think I understand why it is happening but I'm unsure on how to solve it.
Here's a simplified version of what I'm doing;
START TRANSACTION;
-- check if user has already signed up (returned rows > 0, throw error if so)
SELECT * FROM users WHERE email = 'example#site.com' FOR UPDATE;
-- user has not signed up yet.. create the account.
INSERT INTO users SET ...;
COMMIT;
Now this in itself works fine. However when two parallel request happen, a deadlock is made because the transaction will both create a FOR UPDATE lock, which is allowed because initially when there is no account signed up yet there are no rows to lock. Atleast, that's what I think is happening.. correct me if I'm wrong.
I'm curious on how I were to fix this, I still want to check whether not an account has registered already so I can show the user a message. Of course the email has a unique constraint but I do not want to rely on that because the auto increment index will increment, even when it violates the constraint.
Also I'm using typeorm, a sample of my code;
public async registerUser(email: string, password: string, displayName?: string) {
const connection = await getConnection();
connection.transaction(async (manager) => {
// First we need to make sure that this email isn't already registered. If
// it has been registered we can throw a simple UserError which will be
// caught by our error handler.
const hasAlreadyRegistered = await this.findUser(email, manager);
if (hasAlreadyRegistered) throw new UserError('Email has already been registered.');
// At last we can create the user, linking him to the previously created
// authentication strategy.
const user = new User();
user.email = email;
user.displayName = displayName || randomBytes(8).toString('hex');
user.strategies = [authentication];
await manager.save(user);
logger.silly('> Created user row.');
return user;
});
}
I have solved this by just checking for the constraint error in the end (per suggestion of #Shadow). It saves me a lot of hassle.
Code
try {
await manager.save(user);
} catch (err: any) {
// Check whether or not this entry violates an unique constraint.
if (err.code === 'ER_DUP_ENTRY') {
throw new UserError('Email has already been registered.');
} else throw err;
}
i am creating a user, with 'email' field so i want to verify whether that email is already exist or not, if exists error must display. i have my code in express js, node js, angular 6 and mysql database and below is the code to create new user
exports.create = (req, res) => {
// Save to MySQL database
let customer = req.body;
Customer.create(customer).then(result => {
// Send created customer to client
res.json(result);
});
};
where should i use if statement in above code
Thanks in advance
I'm thinking the simplest way of solving your problem is making the email column in the database unique. If you try to insert a new user with an already existing email the query will fail.
Another solution would be that you first do a query that looks in the database if an already existing user has the email (from req.body.email). But that would require having two different SQL queries, which I personally would not prefer.
i think you are using Sequelize ORM.
You can do like this
Customer.findOrCreate({
where: {
email: req.body.email,
},
// other datas needs to inserted
defaults: {
name: req.body.name,
username: req.body.username,
},
}).spread((data, created) => {
if (created) {
// your logics
} else {
res.status(400).send(`${req.body.email} already exists.`);
}
});
I am trying to learn node.js so that I can actually get started on working on my personal project. I have been trying to follow the examples in the book "Learning node.js" (by Marc Wandschneider). I, at this point, decided to forgo practicing his example, and go straight into using his code as framework for my practice code.
In my practice code, all I am trying to do is connect to my RDS database (no, I am not using Elastic Beanstalk, btw), and output contents of one of the tables. Seems simple enough, but when I whip up the code for it (based on the book), it seems to attempt connection, but get hung up in the process. This is my code:
var pool = require('generic-pool');
var async = require('async');
var mysql = require('mysql');
var host = "<database-name>.cu8hvhstcity.us-west-2.rds.amazonaws.com",
database = "<database-name>",
user = "<username>",
password = "<someLongPassword>";
var dbClient;
async.waterfall([
// 1. establish connection to database
function (callback) {
console.log("Connecting to database " + database + "...");
dbClient = mysql.createConnection({
host: host,
database: database,
user: user,
password: password,
port: 3306
});
dbClient.connect();
},
// 2. select all from a table (let's go for locations)
function (cb)
{
var query = "SELECT * FROM locations"
console.log("running query \"" + query + "\"...");
dbClient.query(query, cb);
},
function (rows, fields, callback)
{
console.log(fields);
for (var i = 0; i < rows.length; i++)
{
console.log(JSON.stringify(rows, null, '\t'));
}
}
], function (err, results) {
if (err)
{
console.log("An error occurred...");
console.log(err);
}
else
{
console.log("Everything successfully completed!");
}
dbClient.end();
})
This is better than first attempt, when I put a database member to the argument passed to mysql.createConnection(), and it complained that database was unknown. In neither case did either "An error occurred..." nor "Everything successfully completed!" output to the window.
Is there any async stuff going on that is resulting in some kind of non-terminating infinite loop or something? How do I fix this?
The book has an associated GitHub page
NOTE:
Neither my example nor the cited GitHub code make use of that pool variable, so it can simply be commented out. All you need to do to run this yourself is to say npm install async,npm install mysql (as well as creating a dummy RDS database to point to, that contains locations table) before copying, pasting, and running this code.
EDIT:
I fixed the issue with database. I realized that the name of the database used '_', not '-'. Same issue (code hangs) still persists...
I did the following:
In the second function in the array, I needed two parameters, not one.
I fixed thus:function(results, cb)
The third function simply needed to callback(null)
I am trying to update some user data via the admin SDK. I thought this would work
function directoryUpdate(userId, userDept, userLocation, userPhone, userTitle) {
var update = {
organizations:
{
name: "Next Step Living",
title: userTitle,
primary: true,
type: "work",
department: userDept,
location: userLocation
},
phones:
{
value: userPhone,
type: "work",
primary: true,
}
};
update = AdminDirectory.Users.update(update, userId);
Logger.log('User %s updated with result %s.', userId, update)
return true;
}
but it is not updating the organization or phone data on the record. It also does not throw any kind of error.
three questions, what is the proper syntax to do this update, I assume this works like the API update and behaves like an upsert, is that true, and what is the best practice for capturing any errors during the update. I would like to return a false when the update fails and capture that info. Thanks in advance for your help.
Thanks for your question!
This "inspired" me to work out how the update API worked, as I had got as far as retrieving a User object, updating the properties but had not worked out how to persist the data back to Google.
So, here's my prototype code, which appears to work (the objective being to reset the user's password based on entries in a spreadsheet).
It doesn't seem the most elegant code to me, being that there are two round-trips to the Admin API, and we have to post the email address twice, but I guess that is a side-effect of the JSON API.
var emailAddress = userListSheet.getRange(row, 1).getValue();
var password = userListSheet.getRange(row, 2).getValue();
Logger.log('Email: %s, Password: %s', emailAddress, password);
// Reset user's password
var user = AdminDirectory.Users.get(emailAddress);
user.password = password;
if (changePasswordAtNextLogin == 'Yes') {
user.changePasswordAtNextLogin = true;
}
AdminDirectory.Users.update(user, emailAddress);
Figured out the syntax issue. You do need a set of [] around the name value pairs under organization and phones. organizations:[{....}], phones:[{...}]}; and no, at the end of primary: true under phones. Also changed it from an update to a patch but not sure if that was really required;
update = AdminDirectory.Users.patch(update, userId);
And Yes, it did behave like an upsert and modified existing data and added new data just like the API.
Still need to figure out the best way to capture any errors though so if you have any suggestions please post them.
Looks like supplying an invalid email address is a fatal error that can not be caught and dealt with in code. What I did was get all the primary emails out of Google, store them in an array, and validate that the email I was using was in that list prior to running the update. Since everything else is just a string or logical replacement it should not throw any errors so I am confident that the script will not fail. Now all I have to worry about is the time limit.