[Docker] Upgraded images to Ubuntu 18.04 LTS.

[jpds]

Upgraded all Docker base images to 18.04.

I haven't upgraded the GPU ones as those are still marked experimental: https://hub.docker.com/r/nvidia/cuda/

[googlebot]

Thanks for your pull request. It looks like this may be your first contribution to a Google open source project (if not, look below for help). Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

📝 Please visit https://cla.developers.google.com/ to sign.

Once you've signed (or fixed any issues), please reply here (e.g. I signed it!) and we'll verify it.

---

What to do if you already signed the CLA

Individual signers

• It's possible we don't have your GitHub username or you're using a different email address on your commit. Check your existing CLA data and verify that your email is set on your git commits.

Corporate signers

• Your company has a Point of Contact who decides which employees are authorized to participate. Ask your POC to be added to the group of authorized contributors. If you don't know who your Point of Contact is, direct the Google project maintainer to go/cla#troubleshoot (Public version).
• The email used to register you as an authorized contributor must be the email used for the Git commit. Check your existing CLA data and verify that your email is set on your git commits.
• The email used to register you as an authorized contributor must also be attached to your GitHub account.

[jpds]

CLA signed.

[googlebot]

CLAs look good, thanks!

[angerson]

Looks good to me, but @gunan, is this safe to do? I don't know if there are any problems keeping us at 16.04.

[gunan]

I think ubuntu 18.04 has python 3.7?
If yes, i think this will break all our python 3 docker images.

[bhack]

Yes it has python 3.7 default but what are the python 3 docker images that are you talking about?

[bhack]

Check also #17022

[gunan]

All the images that have "py3" suffix here:
https://hub.docker.com/r/tensorflow/tensorflow/tags/

[gunan]

I will close this PR for now.

[jpds]

@gunan Actually, 18.04 LTS has Python 3.6.6:

$ apt-get install python3
$ python3
Python 3.6.6 (default, Sep 12 2018, 18:26:19)

[bhack]

Seems different https://packages.ubuntu.com/bionic/python
https://packages.ubuntu.com/bionic-updates/python3

[bhack]

https://packages.ubuntu.com/bionic-updates/python3.7

[jpds]

@bhack python3 ultimately installs python3.6 which pulls in 3.6.6: https://packages.ubuntu.com/bionic-updates/python3.6

[bhack]

Mhh it is strange cause https://packages.ubuntu.com/bionic-updates/python3 gives 3.6.5 ... But apt wins

[bhack]

So can we reopen this?

[ymodak]

@gunan Friendly ping. Can you please take a look at this PR? Thanks!

[angerson]

Our Docker images are updated nightly as long as that nightly package builds (for the nightly- images) and per-release (for latest- and XX.XX-), so if your deployment is time-critical and you only want to use stable releases, I'd recommend using your own image until these changes are deployed.

[angerson]

Note that the GPU development images were not changed in this PR. nvidia/cuda doesn't provide Ubuntu 18 / cuda 9.0 base images.

[bhack]

@angersson Is TF still default on 9.0? Any plan for 9.2? There was a bug in cuda 9.0 series with "modern" version of GCC>=6. Some Asan flags for cuda like protect_shadow_gap=0 are not avaialbe on gcc 5.x.

[angerson]

@bhack As far as I know, the TF team is working on upgrading directly to CUDA 10. @tfboyd, can you weigh in?

[bhack]

#22706