Roots Discourse

Multiple Projects gulp watch error

I have set up two different projects recently that mirror the example project setup … they are identical except for the actual site.com information and database passwords.

Both projects load in the browser perfectly …

The issue is that one, the most recent one, will not work on gulp watch … when I run gulp watch it says contacting local host and then fails with Safari saying it can’t connect to localhost or Chrome giving me the
ERR_CONNECTION_REFUSED -

[BS] in the terminal continues to say it is watching, but, in fact it isn’t able to

Has anyone experienced this? Do I need to somehow make multiple host entries … used to do this in MampPro but don’t know how to do it in this new setup yet so any and all replies welcome and appreciated~

If you added the last site after you ran vagrant up the first time, you will probably have to run vagrant reload. I suggest you read this thread:

I went through the setup of each client project the same … following the example project instructions all the way … so they were both created in the same order … each is individual and self contained so not sure why I would need to run reload in that instance … the example with Trellis is because he is using one vagrant machine for multiple projects no?

Even with that said though … I did in fact already try that last night … not the re-provision, but, just the reload … I will try the reprovision as well right now …

I was under the impression that it was standard practice to be able to have multiple project machines at any given time but I’m thinking that perhaps browsersync isn’t set up to see multiple hosts at localhost:3000 and it’s retaining only the first entered host name for whatever reason because no matter what I do I can vagrant up my first machine/project and gulp watch just fine but just can’t get the second one to load~ That’s all that seems logical to me but again … I’m so new to all the little bits and bobs in the gooey center of ansible/gulp/vagrant and the playbooks etc. that I am clueless to troubleshoot at any respectable level yet :wink:

My environment looks like this:

~/Projects/clientprojectONE.com/ansible
~/Projects/clientprojectONE.com/site
~/Projects/clientprojectONE.com/Vagrantfile
~/Projects/clientprojectONE.com/.git
~/Projects/clientprojectONE.com/.gitignore
~/Projects/clientprojectONE.com/.vagrant
~/Projects/clientprojectONE.com/.DS_Store
~/Projects/clientprojectONE.com/README.txt
.git is set to push to remote private github repo and my deploys will all be via ansible only … even if I have to deploy production files locally and copy up to my centos server because I don’t have a playbook set up for centos yet due to my meak skills at the moment~

~/Projects/clientprojectTWO.com/ansible
~/Projects/clientprojectTWO.com/site
~/Projects/clientprojectTWO.com/Vagrantfile
~/Projects/clientprojectTWO.com/.git
~/Projects/clientprojectTWO.com/.gitignore
~/Projects/clientprojectTWO.com/.vagrant
~/Projects/clientprojectTWO.com/.DS_Store
.git is set same as above to it’s own private repo but will deploy to DigitalOcean droplet ( have already successfully gone through the whole process on this one )

okay went ahead and ran vagrant provision on Project TWO and still no response from gulp watch … same error … Unable to load the webpage because the server sent no data.

Now Project ONE also gives the same error.

Okay … in digging deeper into finding any differences between the two projects … I’ve discovered that in the site folder of Project TWO there is no “scripts” directory nor any entries for creating salts in the composer.json file … I’ve also just discovered that the manifest.json file in sage itself does not have the local project’s url as it states should be done on the sage repo page …


“config”: {
“devUrl”: “http://project-name.dev
}

The entry is there on ProjectONE manifest.json file and I don’t recall having entered it manually … so, I’m wondering if, perhaps I neglected to run a command somewhere on Project TWO that resulted in these items not being installed/created or if I did something in a different order that resulted in these differences.

Does anyone know how the correct dev url gets inserted automatically into manifest.json?

I’ll wait on destruction for a minute to see if anyone responds here so I can test this … I’d really rather have a deeper understanding and get to the bottom of this than destroy rebuild and just hope for the best :wink:

Okay … I believe I am getting closer to a resolution of what has happened …

It looks like Scott did some work that perhaps sent my already fragile understanding spiraling into oblivion temporarily … heh …

Update: … I ran vagrant reload and vagrant provision on both Projects … as well as composer update just for good measure … entered the dev url in manifest.json and now have two working projects … yay … all done here :wink: