I have just deployed my first production Meteor app, but ran into a pretty significant issue. When I tried to run my app, I get the following error:
/home/hiapp/bundle/programs/server/node_modules/fibers/future.js:173
throw(ex);
^
Error: /home/hiapp/bundle/programs/server/npm/npm-bcrypt/node_modules/bcrypt/build/Release/bcrypt_lib.node: invalid ELF header
at Module.load (module.js:356:32)
at Function.Module._load (module.js:312:12)
at Module.require (module.js:364:17)
at require (module.js:380:17)
at bindings (/home/hiapp/bundle/programs/server/npm/npm-bcrypt/node_modules/bcrypt/node_modules/bindings/bindings.js:74:15)
at Object.<anonymous> (/home/hiapp/bundle/programs/server/npm/npm-bcrypt/node_modules/bcrypt/bcrypt.js:1:97)
at Module._compile (module.js:456:26)
at Object.Module._extensions..js (module.js:474:10)
at Module.load (module.js:356:32)
at Function.Module._load (module.js:312:12)
Through some research, it seems that the issue is that bcrypt is platform-specific, and because my development is on OS X 10.10, but my production server is on Ubuntu 14.04, there-in lies the problem. My work around was to run npm install bcrypt from the programs/server folder, then copy that from programs/server/node_modules/bcrypt to programs/server/npm. This seems like a lot of manual steps.
I am just about to re-deploy my app with some fixes, and I am worried that I am going to have to again perform the same steps. Is there a better way to deploy a Meteor app where you do not run into this issue? My procedure currently is:
Run meteor build
scp resulting tarball to my server (Digital Ocean)
explode the tarball in my app users home directory
ADDED STEPS: recompile bcrypt and copy the directory over as stated above
I am guessing the auto-refresh will be broken as the first time the server tries to load the new libraries it will explode in fiery glory. Would it be better to simply clone the GIT repo on the server and do the build straight from there, or can I use mup or any other tools to help smoothen the deployment process?
There are two ways of getting around this time-consuming problem which are totally dependant on how you deploy.
Manual deploy
If you are manually deploying your app then make sure that you are using node v0.10.361 and only that. Meteor does not work well with node v0.12.x. More specifically, the fibers module causes a lot of problems, it chokes on various errors2.
The following is the procedure that fixed it for me3, 4:
Remove the bcrypt module from the npm directory:
$ cd path_to_your_app/bundle/programs/server
$ rm -rf npm/npm-bcrypt/node_modules/bcrypt/
Install the bcrypt module into the node_modules directory first. This builds bcrypt against the server's OS:
$ npm install bcrypt
Move the newly created bcrypt module into the npm directory:
$ cp -r node_modules/bcrypt npm/npm-bcrypt/node_modules/bcrypt
Finally, restart the app, mogodb and any web server processes such as nginx6, through upstart if you have it configured. Under the heading sanity steps below you can perform if the steps outlined above have not fixed it for you.
Using Meteor Up (mup)
If you are using mup then the process is much easier as other answers on this thread point out. However, there are cases where errors still arise including the invalid ELF header error. Make sure you have the most recent version of mup npm update mup -g.
The first step is to remove any pre-existing app and nodejs package. Mup installs the app into /opt/ and within there you can find your_app and nodejs. Remove them.
Verify that you have the correct node version, 0.10.36 only1 and the following settings, in the mup.json file:
{
...
"setupMongo": true,
"setupNode": true,
"nodeVersion" : "0.10.36",
"setupPhantom": true,
"enableUploadProgressBar": true,
// Application name (No spaces)
"appName": "your_app",
// Location of app (local directory)
"app": ".",
...
}
Run:
$ mup setup
$ mup deploy
Navigate to the server folder on your server and rebuild modules:
$ cd /opt/your_app/app/programs/server
$ npm rebuild
$ npm install
Optional: $ sudo npm update node-gyp -g
Finally, restart the app, mogodb and any web server processes such as nginx6. After the above steps, you may want to look at sanity steps below for more debugging options.
Sanity steps
There are a few other steps you can perform:
Note that apache may also bind itself to the same port that nginx is running on. Look at your app's error logs and your web server's error logs to see if there are any issues. Stop apache via sudo service apache2 stop or brute force close (not recommended) any running processes on port 80.
Then remove or move the apache2 conf file to prevent it from starting up again. All conf files are located in /etc/init or /etc/init.d.
Use $ mup logs -f to view mup logs. To view errors from your app, the end of the file, /var/log/upstart/your_app.log is useful (assuming you have upstart configured).
If using nginx make sure your directories are symlinked.
$ ln -s /etc/nginx/sites-available/your_server_config /etc/nginx/sites-enabled/your_server_config
Moreover, make sure that you have specified a default_server5 in your_server_config.
Check that mongodb can run.
References
Meteor version 1.0.4.1 (released near the end of March 2015) requires that node version 0.10.36 be installed. I recommend using the node version manager, n, to control what version of node is running on your server at any one time. Link.
The dreaded [XXX.XX.XXX.XX] Error: '/opt/your_app/programs/server/node_modules/fibers/bin/linux-x64-v8-3.28/fibers.node' is missing. Try reinstalling 'node-fibers'? error. Link.
The procedure is modified from here: Link.
For your information, I used DigitalOcean's very helpful guide on deploying a meteor app to a server. Link.
Specifying a default_server. Link.
Common processes to restart are: sudo service mongod/nginx/yourapp restart.
I had the exact same issue - I switched to using Meteor Up (MUP) instead and the problem went away. Not to mention its a lot more straightforward!
I followed this tutorial on youtube - https://www.youtube.com/watch?v=WLGdXtZMmiI which covers deploying using MUP to Digital Ocean.
MUP's github page specifically mentions that it handles binary NPM modules - https://github.com/arunoda/meteor-up#binary-npm-module-support
Hope that's some help!
Running npm rebuild at bundle/programs/server/npm/ should take care of rebuilding the modules for the server platform, regardless of the original app being bundled at Linux or MacOS.
I had the same issue and did the following to fix it:
Uninstall nodejs
sudo apt-get remove nodejs
Remove any remaining files
rm -rf ~/.nvm ~/.npm
sudo rm -rf /usr/bin/node /opt/nodejs /usr/lib/node_modules /usr/local/lib/node_modules
Upgrade node version from the default setting in mup.json
"nodeVersion": "0.10.33",
Run mup setup and deploy
mup setup
mup deploy
Related
I'm using CodePipeline to deploy whatever is on master branch of the git to Elastic Beanstalk.
I followed this tutorial to extend the default nginx configuration (specifically the max-body-size): https://medium.com/swlh/using-ebextensions-to-extend-nginx-default-configuration-in-aws-elastic-beanstalk-189b844ab6ad
However, because I'm not using the standard eb deploy command, I dont think the CodePipeline flow is going into the .ebextension directory and doing the things its supposed to do.
Is there a way to use code pipeline (so i can have CI/CD from master) as well as utilize the benefits of .ebextension?
Does this work if you use the eb deploy command directly? If yes, then I would try using the pipeline execution history to find a recent artifact to download and test with the eb deploy command.
If CodePipeline's Elastic Beanstalk Job Worker does not play well with ebextensions, I would consider it completely useless to deploy to Elastic Beanstalk.
I believe there is some problem with the ebextensions themselves. You can investigate the execution in these log files to see if something is going wrong during deployment:
/var/log/eb-activity.log
/var/log/eb-commandprocessor.log
/var/log/eb-version-deployment.log
All the config files under .ebextension will be executed based on the order of precedence while deploying on the Elastic Beanstalk. So, it is doesn't matter whether you are using codepipeline or eb deploy, all the file in ebextension directory will be executed. So, you don't have to worry about that.
Be careful about the platform you're using, since “64bit Amazon Linux 2 v5.0.2" instead of .ebextension you have to use .platform.
Create .platform directory instead of .ebextension
Create the subfolders and the proxy.conf file like in this path .platform/nginx/conf.d/proxy.conf
In proxy.conf write what you need, in case of req body size just client_max_body_size 20M;
I resolved the problem. You need include .ebextension folder in your deploy.
I only copy the dist files, then I need include too:
- .ebextensions/**/*
Example:
## Required mapping. Represents the buildspec version. We recommend that you use 0.2.
version: 0.2
phases:
## install: install dependencies you may need for your build
install:
runtime-versions:
nodejs: 12
commands:
- echo Installing Nest...
- npm install -g #nestjs/cli
## pre_build: final commands to execute before build
pre_build:
commands:
- echo Installing source NPM dependencies...
- npm install
## build: actual build commands
build:
commands:
# Build your app
- echo Build started on `date`
- echo Compiling the Node.js code
- npm run build
## Clean up node_modules to keep only production dependencies
# - npm prune --production
## post_build: finishing touches
post_build:
commands:
- echo Build completed on `date`
# Include only the files required for your application to run.
artifacts:
files:
- dist/**/*
- package.json
- node_modules/**/*
- .ebextensions/**/*
Ande the config file /.ebextensions/.node-settings.config:
option_settings:
aws:elasticbeanstalk:container:nodejs:
NodeCommand: "npm run start:prod"
I know how to package and then deploy meteor application. But recently for one project i'm stuck at an error which i couldn't resolve.
Steps I followed for package and deploy of my meteor app:
1. meteor build package
2. cd package
3. tar -xf inventoryTool.tar.gz
4. cd bundle/programs/server
5. npm install
6. cd ../..
7. PORT=<port> MONGO_URL=mongodb://127.0.0.1:27017/dbName ROOT_URL=http://<ip> node main.js
Here is the log for the error when i run the npm install(STEP 5) command.
Is there anything missing in my execution?. I'm not using the fibers package anywhere in my project. Does anyone have solution to this problem? Thanks in advance.
Why this happens (a lot)?
Your local version of node is v8.9.4. When using the build command, you will export your application and build the code against this exact node version. Your server environment will require this exact version, too.
An excerpt from the custom deployment section of the guide:
Depending on the version of Meteor you are using, you should install
the proper version of node using the appropriate installation process
for your platform. To find out which version of node you should use,
run meteor node -v in the development environment, or check the
.node_version.txt file within the bundle generated by meteor build.
Even if you don't use fibers explicitly it will be required to run your Meteor app on the server correctly.
So what to do?
In order to solve this, you need to
a) ensure that your local version of node exactly matches the version on the server
b) ensure that you build against the server's architecture (see build command)
To install a) the very specific node version on your server you have two options:
Option I. Use n, as described here. However this works only if your server environment uses node and not nodejs (which depends on how you installed nodejs on the server).
II. To install a specific nodejs version from the repositories, you may do the following:
$ cd /tmp
$ wget https://deb.nodesource.com/node_8.x/pool/main/n/nodejs/nodejs_8.9.4-1nodesource1_amd64.deb
$ apt install nodejs_8.9.4-1nodesource1_amd64.deb
If you are not sure, which of both are installed on your server, check node -v and nodejs -v. One of both will return a version. If your npm install still fails, check the error output and if it involves either node or nodejs and install the desired distribution using the options above.
To build b) against the architecture on your server, you should use the --architecture flag in your build command.
I am trying to get my localhost working on my remote (mediatemple) server.
I have bundled it up and have a /myurl.com/bundle folder with the following files.
this folder contains
main.js
npm-debug.log
programs
server
How do I get this to run?
You should take a look in the README inside the bundle folder. Normally everything ist described there to start your app.
Make sure that NODEJS and MONGO is installed on your remote server. This is NOT included in your bundle as well as NODEJS is not present.
If you are running a system like debian or ubuntu normally you can do the installation with
apt-get install nodejs mongo
Make sure, that the nodejs has release v0.10.36 or v0.10.38
node --version
At the README you see the necessary ENV-VARS like MONGO_URL and PORT you need to set to start your meteor app.
If you have running a apache server already the PORT 80 is already blocked, so try PORT=3000 to start your meteor app.
Example:
MONGO_URL='mongodb://localhost:27017/yourapp' ROOT_URL="http://yourhost" PORT=3000 node main.js
If using as above you do not need to export the ENV-VARS before start
Sometime when starting, there are missing NPM – you get fiber errors
In that case
cd programs/server
npm install
and the try start again.
Good luck
Tom
(I'm writing this response assuming that you are not worried about scalability issue, respond in comment if you want to scale your app)
The best option for running a node application, which Meteor application is, is by using forever.
npm install forever
forever start simple-server.js
If you want to figure out how to see the log files and how to stop/restart your service, you can run forever --help to see all the commands.
I'm stuck at the mup setup. Trying to set it up on my Dreamhost server.
Here is what I started with after running mup setup:
Meteor Up: Production Quality Meteor Deployments
------------------------------------------------
Started TaskList: Setup (linux)
[mydomain.com] - Installing Node.js
[mydomain.com] ✘ Installing Node.js: FAILED
-----------------------------------STDERR------------
tty present and no askpass program specified
Sorry, try again.
sudo: no tty present and no askpass program specified
Sorry, try again.
Then I installed node manually on my server, and set the mup file to "setupNode": false. Tried again and got that:
Meteor Up: Production Quality Meteor Deployments
------------------------------------------------
Started TaskList: Setup (linux)
[mydomain.com] - Installing PhantomJS
[mydomain.com] ✘ Installing PhantomJS: FAILED
-----------------------------------STDERR-----------------
tty present and no askpass program specified
Sorry, try again.
sudo: no tty present and no askpass program specified
Sorry, try again.
Finally, I also deactivated the PhantomJS install, tried again to run mup setup and got that:
Meteor Up: Production Quality Meteor Deployments
------------------------------------------------
Started TaskList: Setup (linux)
[mydomain.com] - Setting up Environment
[mydomain.com] ✔ Setting up Environment: SUCCESS
[mudomain.com] - Copying MongoDB configuration
[mydomain.com] ✘ Copying MongoDB configuration: FAILED
-----------------------------------STDERR-----------------
Warning: Permanently added 'mydomain.com,69.163.152.69' (RSA) to the list of known hosts.
scp: /etc/mongodb.conf: Permission denied
Killed by signal 1.
-----------------------------------STDOUT-----------------
----------------------------------------------------------
Completed TaskList: Setup (linux)
I am not sure what to do or try next. Thanks in advance for your help and suggestions.
Its seems like its a Permission problem.
Try with sudo mup setup or whatever name process are you running, initialize with the sudo keyword
From dream Host forums (admin or tutor post).
node.js — which is used by Meteor — causes some weird issues on our shared hosting servers that can trigger this behavior. We're aware of
the issue, but, for various reasons, it's been difficult to fix.
That being said, Meteor won't work on a shared hosting account anyway,
as it runs as a persistent server process, which isn't permitted.
You'd need a DreamHost VPS or dedicated server to run Meteor.
So if you don't have a dreamHost VPS or dedicated server, and you only want to deploy the app give a try to Modulus.io, it works pretty fine with meteor, or use the meteor deploy servers, doc here
This happens because Sudo isn't installed on the target machine.
On my Debian target machine I did apt-get install sudo to resolve this.
For a RedHat flavor (Centos etc) target machine you might do yum install sudo etc.
However mup is supported for Ubuntu only so it's very likely to run into issues with other flavors. You might want to stick with Ubuntu target machines to avoid headaches like these.
I have been working with Meteor apps for a few months. For the first time, anytime I try to run a meteor app, install via (curl https://install.meteor.com | /bin/sh) or run meteor update, I get this error:
module.js:340
throw err;
^
Error: Cannot find module 'fibers'
at Function.Module._resolveFilename (module.js:338:15)
at Function.Module._load (module.js:280:25)
at Module.require (module.js:362:17)
at require (module.js:378:17)
at Object.<anonymous> (/usr/local/meteor/app/meteor/meteor.js:1:75)
at Module._compile (module.js:449:26)
at Object.Module._extensions..js (module.js:467:10)
at Module.load (module.js:356:32)
at Function.Module._load (module.js:312:12)
at Module.runMain (module.js:492:10)
In the past I'd get random js errors and reinstalling would fix, but I cannot even run the installer or update now. Perhaps it is related to 0.6. Thanks for any insight!
This solution worked for me:
Install the latest Node via the graphical installer at http://nodejs.org/ EDIT: should not be necessary.
Install Fibers via sudo npm install fibers#1.0.1 -g per #Akshat EDIT: should not be necessary.
Delete the file /usr/local/bin/meteor
Delete the folder ~/.meteor
curl https://install.meteor.com/ | sh
I had this happen (again) and did more research. I ended up doing these two steps to address it:
Remove Meteor:
rm -rf /usr/local/meteor /usr/local/bin/meteor
Install Meteor:
curl https://install.meteor.com | /bin/sh
I ended up going to my working computer, grabbing the usr/local/bin files and copy/pasting to the non-working computer. May not be ideal, but it got the job done and I am back coding on both machines. Thank you Akshat very much for the diligence on this issue!
Paul
Try restarting your computer/reloading a terminal because the new meteor is installed in a new location too (it looks like yours is still at /usr/local instead of ~/.meteor)
If that doesn't work install fibers with npm install fibers#1.0.0 -g.
I came across this problem too. As it turns out it was a permissions problem with meteor installed on my mac as root. I reinstalled meteor as the logged in user (with files that are in ~/.meteor). Everything then worked fine.
When you bundle / build meteor on one environment (i.e. OS X) and then try to deploy on a different environment (i.e. Ubuntu) it seems that meteor includes the wrong libraries.
I could be wrong in this, in any case I found that uninstalling and re-installing the bcrypt library worked for me (well at least it got me to the point where I get other errors)
Find the directory where npm list bcrypt exists, then type npm remove bcrypt and then npm install bcrypt.
In Meteor V1.2 this is easy to find. In Meteor V1.3 I think it's in .../bundle/programs/server/npm/node_modules/meteor/npm-bcrypt