EB's picture

Greetings,

I have a vague understanding that TKLBAM is supposed to be used with AWS which eventually has a cost.

I'd like to explore the use of otherwise unused usb flash I have around here. I have very little content on my TKL Wordpress to backup but the configuration and appearance was time consuming and I'd like to preserve that as much as possible so that it can be easily restored at a later time if necessary.

Is there a way to do this automatically at certain times? Should the usb storage not be connected to the computer most of the time, to avoid potential for loss from hacking or other intentional destruction via network?

Welcoming thoughts and advice. Thank you.

Forum: 
Jeremy Davis's picture

Yes it's possible to backup to a USB (or any other local filesystem). Although personally, as per your note re "avoid potential for loss from hacking or other intentional destruction via network" it might be better to backup to some alternate network location?

Also regarding cost, if you only have one server to backup, then you could use the "backup free" Hub plan (if you wish to avoid Hub costs, be sure to downgrade to free plans before the 14 day free trial ends). Then you would only have AWS S3 storage costs. Considering that TKLBAM is a "smart backup" (so only stores the files it needs, rather than the whole OS) considering your use case, the backups will likely be tiny. You can also limit the number of full backups the Hub stores, so if your site is relatively static and you only keep a few months of backups, the costs would be minimal.

FWIW AWS S3 costs less than $0.03/GB/month so on a small WordPress site with 3 months of backups, it's likely that your annual costs would be less that $1/year! Although it's important to note that there may be other costs incloved with S3, such as data transfer. Please see the full AWS S3 pricing page for full details (although I'm sure the full price will still be more than reasonable).

If you still wish to persist with using a USB for your backups, then the TKLBAM FAQ may assist, specifically here and here.

It's also worth noting that your USB won't auto mount, you'll need to do that manually. But considering your desire to (possibly) not have it mounted all the time, perhaps that's a feature in your case? Regarding mounting and unmounting of the USB, you could edit the tklbam cron job to also take care of that if you wished, although it might be useful to include a test to ensure that it's mounted by the time that tklbam runs? An alternate path would be to instead create a TKLBAM hook script to mount/unmount the USB. The beauty of this second path is that you could then also have it mount/unmount when you wish to restore from your backup?

EB's picture

Because you are advocating heavily for AWS I will look at it some more. I looked at the AWS link you provided and this https://aws.amazon.com/free/?all-free-tier.sort-by=item.additionalFields.... Clearly much more reading is required.

Thank you

Jeremy Davis's picture

I'm primarily advocating for AWS because it "just works". Also having an offsite backup is definitely a desirable thing IMO.

Other options are totally legitimate, but will require much more engagement to get working.

EB's picture

Ok, I went through the directions. Looks like an anr was created on AWS and copy / pasted into TKL's interface at which point the instructions ended I was presented with what appears to be pay based service on TKL? Is that what I'm supposed to find? Thank you.

Jeremy Davis's picture

Please see the full details here. If you have any further concerns or questions, please do not hesitate to ask.

EB's picture

Ok. Thank you. I understand the creators of TKL would like to get some compensation for their time and efforts. :-)

I'm trying to get Let's Encrypt to work. I've been using TKL on an unused laptop. It's working fine but I want to make things more correct, including getting https working, a good backup solution, etc.

I tried the Let's Encrypt process over the weekend and again this evening. Errors resulted. Wondering if there is a network issue or anything that might be evident to you.

http://1-1.hopto.org

There is mention of errors in the days log file, but that path is not specified.

Images of the errors:

https://photos.app.goo.gl/4BmSkxYVxUPmAgyt9

Thank you again.

Jeremy Davis's picture

Unfortunately, it's a known issue. FWIW, there are a few threads here on the forums which discuss it. The short answer (and the solution) is to update to a newer version of Dehydrated (the Let's Encrypt client we use on the backend), plus a newer version of Confconsole (to support the newer version of Dehydrated). The step-by-step instructions on how to do all that is covered in the Confconsole v1.1.2 release notes.

FWIW, Let's Encrypt changed things which broke Dehydrated. But the updated Dehydrated broke Confconsole in numerous different ways. But after a few false starts, if you follow the above procedure, you should be up and running in now time... :)

EB's picture

Ok, thank you. :-)

 

Will there be an open to upgrade in place to the next major version? If yes, will that hopefully fix Let's Encrypt and allow users to enable https access with relative ease?

 

Will BAM set up and function more easily in the next major version?

 

Using TKL WordPress. Will the new version have a version of PHP that WordPress does not see as outdated?

 

Thank you for all the great work you do.

 

Jeremy Davis's picture

Will there be an open to upgrade in place to the next major version? If yes, will that hopefully fix Let's Encrypt and allow users to enable https access with relative ease?

As per previous versions you will be able to do a Debian-style "inplace upgrade" or migrate your data via TKLBAM.

Re Let's Encrypt, yes it will "just work" out of the box (at least until Let's Encrypt next break things anyway...).

Although did you have troubles getting it working on v15.x with the updated Confconsole? With the new version of Confconsole installed (and the updated Dehydrated) Let's Encrypt should "just work" already?!

Having said that, moving forward, I will have full access to the TurnKey Buster (the codename for v16.x) apt repositories. So I will be able to push updated packages when need be. That will make fixes for similar issues in the future much easier to install...

Will BAM set up and function more easily in the next major version?

Firstly, I'd be interested if you could clarify exactly what you mean by "set up and function more easily"? Could you give some concrete examples of the difficulties that exist as you see them? What it should do that it doesn't? How do you think it should work, that it doesn't?

Regardless, it will essentially be the same as in v15.x. We had hoped to do some significant work on it for v16.0, but it got out of control and became a much bigger job than we had anticipated. Considering how far behind schedule we are already with v16.0, we decided to scrap that plan and push ahead with TKLBAM as it is (FWIW all the other TurnKey software has been ported to python3 and had significant work done on it - TKLBAM is the biggest and most complex single piece of software that we have).

We do still hope to do some significant work on TKLBAM within the lifetime of v16.x (hence why I ask the questions above). I'm not 100% sure how it will go or exactly when it will happen, but we'll be looking for testers when the time comes. I've also been considering how we might encourage as many users as possible to assist us with testing. I'm sure we'll get quite a few happy to help out, but I'm thinking perhaps we could give away some Hub credits or something similar as an extra incentive?!

I'm open to further thoughts and feedback regarding any of that. So please don't hesitate to share.

Using TKL WordPress. Will the new version have a version of PHP that WordPress does not see as outdated?

Yes, it will come with PHP 7.3. So that should keep WordPress (and other PHP apps that whinge, such as Joomla, etc) happy for another few years. Although unfortunately, it will be inevitable that the current situation will reoccur once PHP upstream (i.e. the PHP developers) end support for PHP 7.3. As per PHP 7.0 in v15.x though, it will continue to get security updates for the life of the Buster LTS (sometime mid 2024).

We could install PHP from an alternate source (instead of Debian) and thus give a newer version. However, there are implications for that (e.g. no auto security updates) so we're not sure that we'd want that by default. However, if we can make that easier, then that's certainly something we'd like to look at.

Thank you for all the great work you do.

Thank you for your kind words... We do our best and hope that yourself and other users do find it relevant and useful. Hopefully TurnKey is also constantly evolving and getting better, although unfortunately, that progress is often much slower than I'd like...

As a bit of an update, over the last few days I've just published all of our packages to our 'buster-testing" apt repository. I still have some work to do on the bootstrap (a minimal collection of operating system packages that we use as the base when building appliances). But I have just done a test build of Core v16.0RC1. And it completed successfully! Yay!

I still need to test it to ensure that it all works, but assuming that goes to plan, then my next step will be to finish tidying up the bootstrap building process. As soon as I've done that, we can publish our bootstrap and I can build Core v16.0RC1 for real! Getting close now... :)

EB's picture

Jeremy,

 

Thanks for the information regarding your challenges and intentions. Congratulations on your recent success with Core v16.0RC1.

 

I have put together something on a tkl wordpress instance that I'd like to backup in the most TKL way possible, using BAM. I'd be concerned about making a change that will cause problems for my wordpress project or future OS / TKL updates that could fix these problems.

 

I will assume that the new major version will be coming in weeks or months rather than years. I hope to follow standardized instructions to make that upgrade and that may not be possible the more changes I make now.

 

Once there is a recoverable backup and I have some level of confidence that I could use it effectively then I could be more adventurous. Perhaps I should put a separate installation on a virtual machine and experiment.

 

My work keeps me super busy and I have 1-2 hours to myself at the end of the day. I'd rather have fun and just focus on the community building project rather than starting over or learning how to fix problems I created. :-)

 

Thank you,

 

Eric

Add new comment