MacBook Pro for Developer

[juro]

It seems my 2015 MacBook Pro is finally dying, so it's time get a new one. As I do a lot of development using virtual machines, I'm wondering whether I need 32GB, or whether I can get away with 16. I'll have to upgrade to 1TB SSD anyway ... Any developers want to chime in?

 

[SauRoNZA]

It seems my 2015 MacBook Pro is finally dying, so it's time get a new one. As I do a lot of development using virtual machines, I'm wondering whether I need 32GB, or whether I can get away with 16. I'll have to upgrade to 1TB SSD anyway ... Any developers want to chime in?

Well virtual machines are going to be a problem…so you are going to have to change how you work.

As such, no you don’t need 32GB. Don’t need 1TB either.

 

[juro]

Well virtual machines are going to be a problem…so you are going to have to change how you work.

As such, no you don’t need 32GB. Don’t need 1TB either.

Thanks for your reply. "Going to be a problem" ... in what sense?

 

[SauRoNZA]

Thanks for your reply. "Going to be a problem" ... in what sense?

You no longer get Intel MacBooks, they are all M1’s now and those don’t do VM’s.

Or rather the other way around most VM’s don’t support M1.

 

[juro]

It seems like VMWare has a compatible version out: https://osxdaily.com/2022/10/22/you-can-now-run-virtualbox-on-apple-silicon-m1-m2/ (currently in developer preview though).

Thank you for your feedback.

 

[SauRoNZA]

It seems like VMWare has a compatible version out: https://osxdaily.com/2022/10/22/you-can-now-run-virtualbox-on-apple-silicon-m1-m2/ (currently in developer preview though).

Thank you for your feedback.

Yes but nothing run inside it.

At least not on any capacity you could work with.

Why the need for VM’s to develop though?

 

[juro]

I maintain various projects across different stacks (e.g. PHP, Python, Angular, React, PostgreSQL, Mysql, MariaDB, etc.) that depend on various combinations of said stacks. for testing I run these in (as light weight as possible) VMs using Docker or Vagrant.

 

[SauRoNZA]

I maintain various projects across different stacks (e.g. PHP, Python, Angular, React, PostgreSQL, Mysql, MariaDB, etc.) that depend on various combinations of said stacks. for testing I run these in (as light weight as possible) VMs using Docker or Vagrant.

So why the need for VM’s at all if you have Docker?

 

[lkpat]

So why the need for VM’s at all if you have Docker?

It's not bad practice to properly sandbox multiple client environments. I can't believe toy can't run a VM on an M series chip?

 

[SauRoNZA]

It's not bad practice to properly sandbox multiple client environments. I can't believe toy can't run a VM on an M series chip?

I mean it depends on the VM.

Now that we have more information the Linux based ones will likely be far less of an issue, it’s mostly Windows that is a problem and usually where the term VM comes into play.

If it’s all in Linux I would just use Docker which is perfectly sandboxed.

 

[Johnatan56]

It's not bad practice to properly sandbox multiple client environments. I can't believe toy can't run a VM on an M series chip?

Docker is sandboxed, variables are set via environment that is container specific.

Vagrant is VM set-up though, so not all is worth migrating if rare.

Macbook's ARM will not work for well for it for legacy stuff, other team I am working with took macbooks and even with containerized, certain stuff is basically dead/will not be ported to ARM, so no support even if docker, they have remote build servers (dumb imho, good as possible, bad as workflow if just quick build).

If you have legacy in your workflow, I would not consider an ARM device currently.

 

[Virex]

It seems like VMWare has a compatible version out: https://osxdaily.com/2022/10/22/you-can-now-run-virtualbox-on-apple-silicon-m1-m2/ (currently in developer preview though).

Thank you for your feedback.

You can try UTM. Last time I tried it I could get Ubuntu working, but that was last year. I assume that it's gotten a lot better since then. It's based on QEMU.

UTM

Securely run operating systems on your Mac

mac.getutm.app

 

[lkpat]

Docker is sandboxed, variables are set via environment that is container specific.

Vagrant is VM set-up though, so not all is worth migrating if rare.

Macbook's ARM will not work for well for it for legacy stuff, other team I am working with took macbooks and even with containerized, certain stuff is basically dead/will not be ported to ARM, so no support even if docker, they have remote build servers (dumb imho, good as possible, bad as workflow if just quick build).

If you have legacy in your workflow, I would not consider an ARM device currently.

The trouble is you can't work with clients who use PowerBI. You then need another PC just for those.

It does seem daft that as a professional you have to maintain two machines.

 

[juro]

Docker on Mac works relatively well, however I have found major performance issues if you are running a large stack with multiple machines. Apparently it's because Docker cannot talk to the file system directly, everything is rooted through the Mac filesystem, which slows it down a lot. One way that can help is to buy Parallels and there was a long blog post about that somewhere.

And here we all though containerizing things would make thing sooooo much easier ...