Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Is it safe to upgrade to ubuntu 16.10 #1660

Closed
paladox opened this issue Feb 2, 2017 · 11 comments
Closed

Is it safe to upgrade to ubuntu 16.10 #1660

paladox opened this issue Feb 2, 2017 · 11 comments

Comments

@paladox
Copy link

paladox commented Feb 2, 2017

Hi, I'm wondering is it safe to upgrade to 16.10 please or will that cause problems. I am asking since another LTS release won't happen for another year.

@benhillis
Copy link
Member

We have not been testing 16.10 internally so you will likely run into some issues.

@paladox
Copy link
Author

paladox commented Feb 2, 2017

Oh, thanks for replying.

@therealkenc
Copy link
Collaborator

16.10 primarily suffers from #1005 as of this writing.

@fpqc
Copy link

fpqc commented Feb 2, 2017

You will also probably run into #555 if you go all the way to 17.04 (just mentioning for completeness) because of the strace update. But yeah, #1005 is the big blocker on a lot of stuff, since they seem to have switched to using it in a ton of GNU devtools.

@stehufntdev
Copy link
Collaborator

@therealkenc, is the failure on 16.10 the same clone call with CLONE_VM|CLONE_VFORK|SIGCHLD? If so do you have a strace of an example failure on 16.10 that we can review?

@therealkenc
Copy link
Collaborator

therealkenc commented Feb 3, 2017

Correct that's the pattern, not the namespace stuff (which got subsumed in there).

I believe, but haven't proved conclusively, it's a glibc thing. For example clang also gets the same error on 16.10 (message) but not on 16.04.

I've updated the test case in #1005 so it actually compiles clean with recent gcc and included the strace. I didn't include the strace at the time because it is just:

[...blah blah blah]
clone(child_stack=0xacef60, flags=CLONE_VM|CLONE_VFORK|SIGCHLD) = -1 EINVAL (Invalid argument)
[...blah blah, not working]

You don't need to install 16.10 to run the test case because I'm deliberately calling clone with the pattern directly.

@stehufntdev
Copy link
Collaborator

Thanks for the follow-up! I was able to get a repro of this on 16.10 and updated the code to handle clone vfork (i.e. CLONE_VM|CLONE_VFORK|SIGCHLD) the same as vfork. As part of this change we realized we have some differences in vfork and those will be patched up later.

@paladox
Copy link
Author

paladox commented Feb 3, 2017

@stehufntdev thank you :)

@therealkenc
Copy link
Collaborator

therealkenc commented Mar 6, 2017

I tried Yakkety (16.10) today and all appears well so far on 15046. The apt-get dist-upgrade with an updated sources.list was utterly painless. I thought strace might break because of #555 (also reported in #1578), but the strace version shipped with Yakkety is 4.12, and strace won't fail on #555 until version 4.15. I'll post again if I come across anything glaring, but I think you can put a fork in this issue because it looks done.

[edit] One tip I forgot to mention for anyone trying this at home. You want to make sure the system dbus is active before running apt-get upgrade, because some package scripts expect it to be live (which is always the case on Real Ubuntu). Dbus started working quite a while back now with AF_UNIX socket fixes #706 (message). Run:

sudo service dbus start

This advice holds for vanilla 16.04 Xenial that ships with WSL too by the way.

@stehufntdev
Copy link
Collaborator

@therealkenc thanks for the update! Closing this out as fixed for insider builds.

@therealkenc
Copy link
Collaborator

therealkenc commented Apr 15, 2017

I suppose I should follow up on this one since it was marked as fixedininsiderbuilds at my behest; which, it technically was at the time due to #1005 being addressed. Unfortunately as of 16176, upgrading to Ubuntu 16.10 is "not safe" per issue title, due to #1878 and friends. I don't think it is necessarily worth re-opening this issue (we have plenty), but it bears clarification in case anyone lands here in a tracker search.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

5 participants