Terra-fork

Okay, so I swiped the headline from the video, but it’s pretty much a given.

Further to my previous blog entry about Hashicorp’s decision to change Terraform’s license, I mentioned how the OpenTF Initiative put a choice to Hashicorp: change the license back or we will fork Terraform.

Unsurprisingly, Hashicorp didn’t comply, so now Terraform has (or will be) forked as OpenTF at https://github.com/opentffoundation

This story has been covered in several places including The Register and several videos on YouTube including a nice short summary by Fireship

If the fork goes ahead, I’m curious as to whether Hashicorp sees it as a competition and goes after the foundation, but if it does, what little reputation Hashicorp has left in the OSS community will be nigh on destroyed.

Meanwhile, I’m going to continue playing with Pulumi in case the OpenTF doesn’t work, gets blocked or if we need another alternative.

Google Domains and Terraform

Two major updates recently

Firstly, as suspected, I finally got the notification saying that Google Domains’ registrations would be acquired by SquareSpace so all my registrations would be transferred over them should I not do anything. Obviously, I didn’t want that, so I transferred them over to AWS (Route 53) so now my domains are registered with AWS, but are DNS managed by Cloud DNS. I had some weirdness when trying to migrate all my domains in bulk, with respect to the auth code not being accepted on one of the domains when doing the bulk migration, but it was accepted when I did the migration on that one domain alone, so… go figure.

Next update is Terraform. In case you didn’t know, Hashicorp has changed the Terraform license and essentially made it no longer open source. This behaviour is similar to what Red Hat did with its RHEL offering and the backlash is just as bad.

Immediately I knew someone would fork it, and already, there’s the OpenTF Initiative and this is the key part:

Our request to HashiCorp: switch Terraform back to an open source license.

We ask HashiCorp to do the right thing by the community: instead of going forward with the BUSL license change, switch Terraform back to a truly open source license, and commit to keeping it that way forever going forward. That way, instead of fracturing the community, we end up with a single, impartial, reliable home for Terraform where the whole community can unite to keep building this amazing ecosystem.

Our fallback plan: fork Terraform into a foundation.

If HashiCorp is unwilling to switch Terraform back to an open source license, we propose to fork the legacy MPL-licensed Terraform and maintain the fork in the foundation. This is similar to how Linux and Kubernetes are managed by foundations (the Linux Foundation and the Cloud Native Computing Foundation, respectively), which are run by multiple companies, ensuring the tool stays truly open source and neutral, and not at the whim of any one company.

OpenTF Initiative (https://opentf.org/)

Essentially, make Terraform open source again, or a fork from the MPL version will be made and maintained separately from Hashicorp’s version. This will essentially lead to two, potentially diverging versions of Terraform, one BUSL and one MPL licensed

I’m already looking at alternatives and the two currently that I’m looking at are Ansible and Pulumi

Ansible I’ve had experience in , but there’s two main issues with it:

  • It’s underused compared to Terraform and the providers are woefully undersupported and undermaintained
  • It’s Red Hat

Pulumi I’ve heard lots of good things about, but its a new technology, and I don’t know if it is can “Import” existing infrastructure.

Guess a “spike” is worth doing for it.

Updates

It’s been quite a long time since I did any updates on this blog so a couple of updates are in order

House

I’ve now been living in the new place for just over a year. Generally everything is good, we’ve started putting in a lawn and currently letting it get its roots in before we try to cut it

Twitter

Oh, boy, what an absolute train wreck. I’ve tolerated Elon’s presence at Twitter because most of the stuff he did wasn’t far off what Jack was doing prior. But cutting off all third party clients, forcing everyone onto the new TweetDeck (which likely will be paywalled too) has literally driven users away. Twitter has been losing users and revenue constantly, and it is totally unsurprising.

I’ve disabled my two TweetDeck profiles (a professional one, and a casual one) from Ferdium. But I doubt I will be going back any time soon.

Red Hat

Another train wreck of a situation. Red Hat’s decision to first kill CentOS’s stability, then for RHEL sources to only be accessible behind a subscription has pissed of a lot of users, even if it’s not strictly speaking against license.

The only one left in its sights will be Fedora so that leads me onto the next update

Manjaro & Archlinux

I’ve been tinkering with Manjaro more and more lately, with its rolling release schedule meaning I never need to upgrade from a major version to another major version.

It’s downside I’m finding is that some packages, especially those in the AUR are essentially “compile from source” packages which does the build on your machine during the install. This can take a varying amount of time depending on the code. With my RSS reader of choice: QuiteRSS, this build takes a mind numbing 2.5 hours to do, even on a high spec machine.

That’s where I found out about setting up your own Arch repo. I’ve been tinkering with that, setting it up on GCP and fronted by a CDN. This works pretty well, but I still need to find how I can do scheduled builds to keep that up to date, but it looks like I’ll be switching to Manjaro at some point in the near future. Sound works fine, using lyncolnmd‘s work.

Another downside with Manjaro, however is that its btrfs filesystem, my home directory backup, and CloneZilla don’t seem to want to work well together

WordPress & Domains

One final update, I will likely be transferring out the blenderfox.com domain out of WordPress, while I had this registered as part of the blog, I’m finding it much harder to maintain this domain using WordPress’s very limited DNS management tools. I will likely transfer it out to Google Domains, even though there’s talk of them shuttering that service. Secondary service would be AWS.

I booted Linux 292,612 times

rich's avatarRichard WM Jones

And it only took 21 hours.

Linux 6.4 has a bug where it hangs on boot, but probably only 1 in 1000 boots (and rarer if using Intel hardware for some reason). It’s surprising to me that no one has noticed this, but I certainly did because our nbdkit tests which use libguestfs were randomly hanging, always at the same place early in booting the libguestfs qemu appliance:

[    0.070120] Freeing SMP alternatives memory: 48K

So to bisect this I had to run guestfish in a loop until it either hangs or doesn’t. How many times? I chose 10,000 boots as a good threshold. To make this easier I wrote a test harness which uses up to 8 threads and parses the output to detect the hang.

After a painful bisection between v6.0 and v6.4-rc6 which took many days I found the culprit, a regression in the printk time…

View original post 23 more words

Citymapper

I paid for the premium version of Citymapper primarily because I need to use their “speak directions” option while on public transport. Something that used to be free, and then they put it behind a paywall.

Now they only went and made it FREAKING FREE again, after I’d gone premium and paid for the year >_<

https://citymapper.com/news/2589/citymapper-club-features-are-now-available-to-all

So I’ve terminated the premium subscription, but I’m still premium until the end of this annual cycle, which ends in November

Coronation and Linux

This weekend was the coronation of King Charles III, the first new monarch for most of the people in the UK. Some may have been alive for the late Queen Elizabeth II’s coronation but the are few now.

I was working coronation support cover on the Saturday from around 6:30am to about 3pm. Fortunately no major issues. One alert towards the tail end of the day, but that was about it.

Sunday were the street parties. Our local one was relatively small compared to the one during the Jubilee celebrations. Was great for the local kids though, and the weather was perfect for it.

Monday is a Bank Holiday so no-one was working in the London office, although the other offices were still working.

I chose today to retry doing my Manjaro file copy, and again it failed with checksum errors when I tried to back it up (even the btrfs check didn’t manage to fix it)

I guess I’ll have to restart my Manjaro attempts and not use btrfs — probably return to using ext4 and lvm.

Fedora

The final issue with the post-upgrade is now fixed (dependency issue) and it self-resolved by a new package update that came through this morning

I did fine one new issue — a new kernel dropped and installed successfully, but did the same thing as the upgrade and did not generate an initrd line, but did generate the initrd line once I regenerated the grub configs.

Fedora

Fixed the VirtualBox issue now.

Looks like on VB6 you had to add the current user to the vboxsf group.

But on VB7, you have to add them to the vboxusers group instead.

Fedora

Managed to fix the issue mentioned in my previous post.

Looks like the system upgrade messed up generating the config during the system upgrade process.

The first time, with the blockdev error and second time, it was missing an initrd directive.

So I booted into the previous installation (which was still available as I made the grub menu visible as part of my Ansible playbook)

Inside the old installation, I just regenerated the grub configs:

sudo grub2-mkconfig -o /etc/grub2.cfg
sudo grub2-mkconfig -o /etc/grub2-efi.cfg

And this readded the initrd line and now I am able to boot into F38. Everything still works, I have some errors with VirtualBox and qemu that I will need to fix at some point:

  • VirtualBox has an error when you start it saying “Unable to enumerate USB”
  • Qemu has a dependency error (must have changed between F36 and F38) so I’ll need to probably remove and reinstall that.

Fedora, Manjaro, WordPress & Twitter

So, quite a few things have been happening recently.

I tried to go the next step on Manjaro and copy my home directory across ready for me to run the Ansible playbook to do the setups (I’ve been testing it would work via VirtualBox on my Windows laptop)

After waiting ages for the file copy, when I went to do the backup, there was a btrfs checksum error so I’ll have to try again some other time.

Separately, I got a message on an issue I raised at Fedora’s bugzilla the Fedora 36 (the version of Fedora I raised the bug on), was coming up to EOL so I should look to move away from it or upgrade. So I decided to try doing the upgrade.

The upgrade completed without error, but first boot after the upgrade hung with a weird Kernel Panic error (I have posted this on the Fedora Forums)

So I rolled my laptop back to preupgrade state for now.

I suspect I may need to end up doing a clean install, so I’ve prepared a btrfs and a ext4 lvm installation and have imaged those in preparation.

Finally, WordPress posted that they’re stopping Autosharing to Twitter because of Twitter’s Idiot-in-Chief screwing up the API usage.

Still, they have mentioned they will be adding autosharing to Instagram and Mastodon instead, and that helps me, since I’m slowly moving what little Twitter presence I had onto Mastodon anyway.

My Twitter has been disconnected from my WordPress and I’ve revoked its access from Twitter’s side anyway.