Cloudron post install questions

ejarvis

I've gone through the installation process successfully, however, when it attempts to show the page content on https://my.domain.com, it briefly displays the screen but then forwards to a standard 404 error.

I had temporarily forwarded ports 80 and 443 to the host private IP, received the same response.

Pinged my.domain.com and verified that it is indeed forwarding to my public IP address via Cloudflare.

Can the default ports be assigned to non-standard ports?

Any help with diagnosing and correcting the 404 issue would be great, thanks in advance.

nebulon

if you briefly see the page, it might mean that some script redirects you. Can you just curl your dashboard domain and see what is actually delivered?

To be on the same page, you were successfully going through the Cloudron DNS and setup screens after installation by visiting your IP ?

ejarvis

@nebulon said in Cloudron.io post install questions:

if you briefly see the page, it might mean that some script redirects you. Can you just curl your dashboard domain and see what is actually delivered?

When I issue: curl https://my.domain.com/setup.html I receive the following response:
curl: (7) Failed to connect to my.domain.com port 443 after 12 ms: Bad access

For now, port 443 is forwarding to the correct private IP address where the server is hosted.

To be on the same page, you were successfully going through the Cloudron DNS and setup screens after installation by visiting your IP ?

Yes, I had completed adding the Global API key that I retrieved from Cloudflare and entered my email address associated with that account.

It went through its process and went to the second page telling me that it was retrieving the certificate for 443 access.

Shortly after, a brief glimpse at the next setup screen (before I can access your app store) and then forwards/resolves to a 404.

nebulon

@ejarvis to rule out some error cases, can you disable the proxying for that domain in cloudflare and check if this may be the issue?

ejarvis

@nebulon I tried that actually and was able to get it connected. However, now I have an app that is stuck on 'installing' and won't let me stop the task, or remove the app (or go through the steps that requires placing the app in a recovery mode).

Is there a way to completely wipe my slate so that I can try to install an app again?

BrutalBirdie

@ejarvis
Might be connected to:
https://forum.cloudron.io/topic/8101/fix-for-kernel-bug-in-ubuntu-20-04-causing-various-issues

girish

@ejarvis said in Cloudron post install questions:

Is there a way to completely wipe my slate so that I can try to install an app again?

You can just reboot the server. That should make it unstuck if this is related to the Ubuntu kernel issue. Can you please try that?

ejarvis

@BrutalBirdie I'm checking this now, thanks for the link.

ejarvis

@girish said in Cloudron post install questions:

@ejarvis said in Cloudron post install questions:

Is there a way to completely wipe my slate so that I can try to install an app again?

You can just reboot the server. That should make it unstuck if this is related to the Ubuntu kernel issue. Can you please try that?

I've tried that. First rule of troubleshooting, eh?

ejarvis

@BrutalBirdie Applied that via SSH, everything went smoothly. However, the app is still stuck in 'installing' mode and won't let me stop or remove it.

Guacamole
^^screenshot

girish

@ejarvis Ah, the "Queued" usually either means the platform is still starting or the backup is running. For the latter, you can check in the Backups view. For the former, can you see /home/yellowtent/platformdata/logs/box.log ? Does it show some error or goes on about platform lock not released?

ejarvis

@girish said in Cloudron post install questions:

@ejarvis Ah, the "Queued" usually either means the platform is still starting or the backup is running. For the latter, you can check in the Backups view. For the former, can you see /home/yellowtent/platformdata/logs/box.log ? Does it show some error or goes on about platform lock not released?

It does say that there is a conflict and that the task is locked.

Box POST /api/v1/apps/cf294eb7-14c1-4f18-813a-14f9b2cb6463/exec 409 Conflict App not installed or running 465.399 ms>
2022-12-01T01:06:16.237Z box:locker Lock unreleased platform_start
Box POST /api/v1/apps/cf294eb7-14c1-4f18-813a-14f9b2cb6463/restart 409 Conflict Locked by task 3 : pending_install />

On the bright side, the log files are very verbose and relatively detailed, with nearly everything that we need to know. Is there something else that I can verify for you?

girish

@ejarvis said in Cloudron post install questions:

2022-12-01T01:06:16.237Z box:locker Lock unreleased platform_start

If you keep scrolling up, you should find some error. One of the services is not starting. Which version are you on?

ejarvis

@girish You mean this?

Box POST /api/v1/tasks/3/stop 409 Conflict task is not active 816.625 ms - 61
Box GET /api/v1/apps/cf294eb7-14c1-4f18-813a-14f9b2cb6463/files/ 404 Not Found Unable to find container sftp 401.543>
Box GET /api/v1/apps/cf294eb7-14c1-4f18-813a-14f9b2cb6463/files/ 404 Not Found Unable to find container sftp 701.603>
Box GET /api/v1/apps/cf294eb7-14c1-4f18-813a-14f9b2cb6463/files/ 404 Not Found Unable to find container sftp 628.086

Top header says Cloudron 7.3.2

girish

@ejarvis no, there should be some other error. You can also do systemctl restart box and it should give some docker error. Otherwise, if you contact us on support@cloudron.io, i can take a look.

ejarvis

@girish Let me do a restart and will see what I can see. If that fails, will contact you at that address.

ejarvis

@girish Well, the result is interesting:

==== AUTHENTICATING FOR org.freedesktop.systemd1.manage-units ===
Authentication is required to restart 'box.service'.
Authenticating as: Edwin Jarvis (jarvis)
Password:
==== AUTHENTICATION COMPLETE ===

Ummm...

BrutalBirdie

@ejarvis
Did you install Cloudron on your Ubuntu Desktop Computer?
It looks like you are running as user jarvis who does not have permission to execute that command.
So you should use sudo systemctl restart box.service

But please answer first on what systems you installed Cloudron.
Like very specific, tell us all.

subven

@ejarvis said in Cloudron post install questions:

==== AUTHENTICATING FOR org.freedesktop.systemd1.manage-units ===
~~
Authenticating as: Edwin Jarvis (jarvis)

Looks like CentOS Desktop to me.

@ejarvis please post the output of cat /etc/*-release If you're not on Ubuntu 20.04 or 22.04 we won't help you any further and I would advice you to follow the Cloudron installation docs.

ejarvis

@subven said in Cloudron post install questions:

@ejarvis said in Cloudron post install questions:

==== AUTHENTICATING FOR org.freedesktop.systemd1.manage-units ===
~~
Authenticating as: Edwin Jarvis (jarvis)

Looks like CentOS Desktop to me.

@ejarvis please post the output of cat /etc/*-release If you're not on Ubuntu 20.04 or 22.04 we won't help you any further and I would advice you to follow the Cloudron installation docs.

Umm... no?

NAME="Ubuntu"
VERSION="20.04.5 LTS (Focal Fossa)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 20.04.5 LTS"
VERSION_ID="20.04"
HOME_URL="https://www.ubuntu.com/"
SUPPORT_URL="https://help.ubuntu.com/"
BUG_REPORT_URL="https://bugs.launchpad.net/ubuntu/"
PRIVACY_POLICY_URL="https://www.ubuntu.com/legal/terms-and-policies/privacy-policy"
VERSION_CODENAME=focal
UBUNTU_CODENAME=focal

I installed a minimal XFCE4 desktop environment so that I could VNC into the server as needed.

Do you think this is part of the issue?

ejarvis

@subven said in Cloudron post install questions:
If you're not on Ubuntu 20.04 or 22.04 we won't help you any further and I would advice you to follow the Cloudron installation docs.

This statement bothers me, since the person that was helping me was directly from the source, and you haven't helped at all yet?

For what it's worth, I did follow the cloudron install docs that you linked. Prior to that, the server has been running on that limited desktop environment.

If I need to nuke it all and do it again, it's not that much of a hassle.

EDIT: Nuking it now.

ejarvis

@subven said in Cloudron post install questions:

@ejarvis said in Cloudron post install questions:

==== AUTHENTICATING FOR org.freedesktop.systemd1.manage-units ===
~~
Authenticating as: Edwin Jarvis (jarvis)

Looks like CentOS Desktop to me.

@ejarvis please post the output of cat /etc/*-release If you're not on Ubuntu 20.04 or 22.04 we won't help you any further and I would advice you to follow the Cloudron installation docs.

Also:

ls -l /usr/share/xsessions
total 12
-rw-r--r-- 1 root root 303 Mar 26 2020 ubuntu.desktop
-rw-r--r-- 1 root root 6273 Mar 30 2020 xfce.desktop

ejarvis

@BrutalBirdie said in Cloudron post install questions:

@ejarvis
Did you install Cloudron on your Ubuntu Desktop Computer?
It looks like you are running as user jarvis who does not have permission to execute that command.
So you should use sudo systemctl restart box.service

But please answer first on what systems you installed Cloudron.
Like very specific, tell us all.

I do notice that there are two users listed in the server now, my own as it was set up when I installed the server, and the Cloudron account that I assume was installed with their script.

The 'jarvis' account is root.

ejarvis

Looking around inside the application, going to services, I can see that neither the MySQL nor PostgreSQL are running.

PostgreSQL at least shows an amber indicator like it's trying to load. Could this also contribute?

BrutalBirdie

@ejarvis said in Cloudron post install questions:

I installed a minimal XFCE4 desktop environment so that I could VNC into the server as needed.

I am a bit confused with the as needed statement.
You only need a fresh Ubuntu 20.04 / 22.04 server install with ssh access to run the Cloudron install commands.
It's generally not advised to install other software alongside Cloudron because it can lead to many problems.

The fastest and easiest way to come to a working state would be a fresh Ubuntu 20/22 install, ssh into root or the sudo user and then (if sudo switch to root) follow the install.

@ejarvis said in Cloudron post install questions:

The 'jarvis' account is root.

What? No? root is root. . . jarvis might be a sudoer?

The Cloudron install needs to be run with root / sudo.
Amazingly this is not mentioned anywhere @staff I never noticed that until now.

https://docs.cloudron.io/installation/

subven

I also think there should be a warning within the docs to not apt update/upgrade or install other packages after the Cloudron installation. It seems not clear enough that Cloudron is taking over the server and handles all that stuff to some users and warning at the SSH welcome banner is likely to be overlooked.

girish

@BrutalBirdie said in Cloudron post install questions:

The Cloudron install needs to be run with root / sudo.
Amazingly this is not mentioned anywhere @staff I never noticed that until now.

The setup script errors out that you have to run as sudo.

ejarvis

@BrutalBirdie said in Cloudron post install questions:

@ejarvis said in Cloudron post install questions:

I installed a minimal XFCE4 desktop environment so that I could VNC into the server as needed.

I am a bit confused with the as needed statement.
You only need a fresh Ubuntu 20.04 / 22.04 server install with ssh access to run the Cloudron install commands.
It's generally not advised to install other software alongside Cloudron because it can lead to many problems.

The fastest and easiest way to come to a working state would be a fresh Ubuntu 20/22 install, ssh into root or the sudo user and then (if sudo switch to root) follow the install.

@ejarvis said in Cloudron post install questions:

The 'jarvis' account is root.

What? No? root is root. . . jarvis might be a sudoer?

The Cloudron install needs to be run with root / sudo.
Amazingly this is not mentioned anywhere @staff I never noticed that until now.

https://docs.cloudron.io/installation/

You're correct, 'jarvis' is a sudoer and not root, I misspoke.

Before I decided to give Cloudron a try, I added VNC as a means to access it without really giving it another thought. Most, if not all, of our other machines use VNC as a means to support the user when they experience an issue in real-time.

Re-reading what everyone has posted thus far, does lean to the concept that in doing that, could have inadvertently caused some issues.

However, if there is a way to enter a 'safe mode' for the lack of another, more proper phase, we could remove the app that's 'stuck' and try to add it again?

Does such a phase exist? Could it, if it doesn't?

ejarvis

@subven said in Cloudron post install questions:

I also think there should be a warning within the docs to not apt update/upgrade or install other packages after the Cloudron installation. It seems not clear enough that Cloudron is taking over the server and handles all that stuff to some users and warning at the SSH welcome banner is likely to be overlooked.

That's a fair point.

Would it still be an issue if those packages were completely removed from the server?

ejarvis

@girish said in Cloudron post install questions:

@BrutalBirdie said in Cloudron post install questions:

The Cloudron install needs to be run with root / sudo.
Amazingly this is not mentioned anywhere @staff I never noticed that until now.

The setup script errors out that you have to run as sudo.

I vaguely remember elevating with sudo on the installation script, and since it was as extensive as it was, I did other things and let it do its job.

Since I saw somewhere that every night the Cloudron service will handle all updates and such, and with part of the documentation stating that we should 'un-hold' some of the updates that were previously on-hold, I haven't checked yet to see if it helped or hindered yet. Will post more if I see things that I'm not sure how to deal with.

In the meantime, if I reinstalled the OS from a live ISO, will the installation script known to connect with my current account, or do I have to remove everything and start over completely from scratch?

subven

@ejarvis said in Cloudron post install questions:

or do I have to remove everything and start over completely from scratch?

The easiest way is to setup a fresh new server, Install Cloudron and during initial setup, restore from a backup. Be sure to install the same Cloudron version your current server is on. I'm not sure if you also need the same OS version (Ubuntu 20.04 in your case) but I would advice to stick with it as you can upgrade later.

@ejarvis said in Cloudron post install questions:

Would it still be an issue if those packages were completely removed from the server?

Not sure what is the best way to fix a half broken Cloudron installation. AFAIK Cloudron expects explicit apt package versions and installing other packages can easy mess with that. While in theory it is possible, I would avoid it and start with a clean OS as it will 100% work.

@ejarvis said in Cloudron post install questions:

I added VNC as a means to access it without really giving it another thought

~~I think VNC to get basic terminal access should not be the problem.~~ Installing a desktop enviroment however is far more complex in terms of related packages. ~~If you install VNC before Cloudron, I think you should be fine as the setup script will pick the right packages anyway.~~ EDIT: Okay so I digged into the topic and VNC seems to require a DE installed (even for a terminal session) at least if it runs on OS level.

I haven't found anything here in the forums about how others solve that if they don't have VNC from their hoster.

ejarvis

@subven said in Cloudron post install questions:

@ejarvis said in Cloudron post install questions:

or do I have to remove everything and start over completely from scratch?

The easiest way is to setup a fresh new server, Install Cloudron and during initial setup, restore from a backup. Be sure to install the same Cloudron version your current server is on. I'm not sure if you also need the same OS version (Ubuntu 20.04 in your case) but I would advice to stick with it as you can upgrade later.

@ejarvis said in Cloudron post install questions:

Would it still be an issue if those packages were completely removed from the server?

Not sure what is the best way to fix a half broken Cloudron installation. AFAIK Cloudron expects explicit apt package versions and installing other packages can easy mess with that. While in theory it is possible, I would avoid it and start with a clean OS as it will 100% work.

@ejarvis said in Cloudron post install questions:

I added VNC as a means to access it without really giving it another thought

~~I think VNC to get basic terminal access should not be the problem.~~ Installing a desktop enviroment however is far more complex in terms of related packages. ~~If you install VNC before Cloudron, I think you should be fine as the setup script will pick the right packages anyway.~~ EDIT: Okay so I digged into the topic and VNC seems to require a DE installed (even for a terminal session) at least if it runs on OS level.

I haven't found anything here in the forums about how others solve that if they don't have VNC from their hoster.

It absolutely does, which was why I had installed a minimal DE of XFCE4.

I think that I am going to have to start over and see if (now that I know that Cloudron requires an unaltered installation of Ubuntu Server) that lets me install the app that I wanted to deploy.

girish

@ejarvis yes, don't install ubuntu desktop and such. Maybe installing some minor tools here and there is OK, but bringing in a whole bunch of other packages is completely untested territory.

subven

@girish do you have a solution to install VNC on a Cloudron host?

ejarvis

All:

You're not going to believe this ...

I set out to recreate the Ubuntu server today, simply because I have nothing else too pressing that needed my immediate attention.

It's important to know that this was a fresh install, with the only package that was installed prior to installing Cloudron being net-tools.

I connected everything, and as my luck would have it, attempted to install Guacamole again, and it hangs at the same place.

In addition to that, my screen is also showing that it's "Starting services" and that it could take a while.

EDIT: Correction, the app is showing "Installing - Waiting for Cloudron to initialize' as its status.

I'm starting to wonder if there is something else that is preventing this from happening that's not directly related to Ubuntu or Cloudron.

As an aside, I have a recursive DNS setup that also uses Unbound, and I had set that as the primary DNS IP and nothing else as a secondary or tertiary. Could this be an issue? If so, why does everything else work flawlessly but not an app install?

I also use Docker ran from another server, and finally, the Ubuntu server is run on a virtual machine.

While I find troubleshooting Linux things very challenging and fulfilling, this one is starting to have the opposite effect. Any ideas and help, as always, is greatly appreciated.

ejarvis

Just checked the Services screen:

Cloudron, docker, mail, mysql, nginx, turn, and unbound are all active and 'green'. Everything else is red.

ejarvis

Besides port 80 and 443, are there any other important ports that should be open for communication?

girish

@ejarvis said in Cloudron post install questions:

EDIT: Correction, the app is showing "Installing - Waiting for Cloudron to initialize' as its status.

The issue is that the (internal) services are all not starting up on the server for some reason. Can you email the entire /home/yellowtent/platformdata/logs/box.log to support@cloudron.io ? There's got to be some error there or maybe it is stuck somewhere else which we are not handling.

ejarvis

@girish It looks like it's choking on MySQL for some reason. I'll send you the log as soon as I am able to.

ejarvis

Sorry for the delay, have had internet issues at the office that was preventing me from grabbing the log file and sending it via email.

Now that we're back online I will get that done as soon as possible.

ejarvis

@girish said in Cloudron post install questions:

@ejarvis said in Cloudron post install questions:

EDIT: Correction, the app is showing "Installing - Waiting for Cloudron to initialize' as its status.

The issue is that the (internal) services are all not starting up on the server for some reason. Can you email the entire /home/yellowtent/platformdata/logs/box.log to support@cloudron.io ? There's got to be some error there or maybe it is stuck somewhere else which we are not handling.

Sorry for the delay, apparently our outage also almost caused file degradation on our NAS and that had to be resolved before this could.

I just sent a response to that email address that also includes Johannes, if you don't see it that is where you can request to look at it.

Please let me know what else I can do.

girish

@ejarvis The issue seems to be that mongodb has crashed/not starting up properly. I sent you some instructions via email.

2022-12-08T14:23:14.794Z box:cloudron Startup task at index 2 failed: Network error waiting for mongodb: connect ECONNREFUSED 172.18.0.5:3000 BoxError: Network error waiting for mongodb: connect ECONNREFUSED 172.18.0.5:3000
at /home/yellowtent/box/src/services.js:588:33
    at processTicksAndRejections (node:internal/process/task_queues:96:5)
at async promiseRetry (/home/yellowtent/box/src/promise-retry.js:17:20)
at async waitForContainer (/home/yellowtent/box/src/services.js:583:5)
at async startMongodb (/home/yellowtent/box/src/services.js:1508:9)
at async Object.startServices (/home/yellowtent/box/src/services.js:806:9)
at async Object.start (/home/yellowtent/box/src/platform.js:67:13)
at async onActivated (/home/yellowtent/box/src/cloudron.js:79:5)
at async Array.<anonymous> (/home/yellowtent/box/src/cloudron.js:130:9)

ejarvis

@ejarvis said in Cloudron post install questions:

@girish said in Cloudron post install questions:

@ejarvis said in Cloudron post install questions:

EDIT: Correction, the app is showing "Installing - Waiting for Cloudron to initialize' as its status.

The issue is that the (internal) services are all not starting up on the server for some reason. Can you email the entire /home/yellowtent/platformdata/logs/box.log to support@cloudron.io ? There's got to be some error there or maybe it is stuck somewhere else which we are not handling.

Sorry for the delay, apparently our outage also almost caused file degradation on our NAS and that had to be resolved before this could.

I just sent a response to that email address that also includes Johannes, if you don't see it that is where you can request to look at it.

Please let me know what else I can do.

I performed those steps, the DB is still not starting up. I am going to respond to that email message and attach the logs so that you can see what's happening.

namitazop

can i install cloudron on fedora server

nebulon

@namitazop Cloudron only supports Ubuntu (20 or 22).

namitazop

okay, i'll try it running on ubuntu using virtual machine.. .

Cloudron makes it easy to run web apps like WordPress, Nextcloud, GitLab on your server. Find out more or install now.

Solved Cloudron post install questions