On Thu, Aug 21, 2014 at 4:19 PM, Luca Bruno wrote:
> Try disabling the firewall in the nixos configuration.
>
In trivial.nix, I replaced
networking.firewall.allowedTCPPorts = [ 80 ];
with
networking.firewall.enable = false;
nixops was still unable to determine when the ssh server came up
Try disabling the firewall in the nixos configuration.
About the virtualbox issue, it's possible that the vbox kernel of your
linux box is not compatible with the nixpkgs virtualbox.
On Fri, Aug 22, 2014 at 1:15 AM, Richard Wallace <
rwall...@thewallacepack.net> wrote:
> Hello again,
>
> I fina
Hello again,
I finally got `nixops deploy -d trivial` to create a virtualbox image and
start it - I had to abandon using virtualbox from nixpkgs and instead put
the VBoxManage installed by Arch on my PATH.
During startup, the IP address is determined correctly and then nixops
says, "waiting for S
Neither of those suggestions worked. I'm not sure what is going on. I've
searched google for the error
[nix-shell:~/Development/test/nix]$ VBoxManage startvm
nixops-d397269c-27fb-11e4-96b7-d1c68637124d-webserver --type headless
Waiting for VM "nixops-d397269c-27fb-11e4-96b7-d1c68637124d-webserve
Also note that the VirtualBox installed by nixpkgs doesn't get setuid root
so either do that manually (breaking statelessness a little) or run as
root.
If the version is an issue, you can override the derivation or do a pull
request :-)
Wout.
On Aug 21, 2014 8:29 PM, "Richard Wallace"
wrote:
>
Grr. Seems I spoke too soon. Looks like the problem is that the VirtualBox
installed via the Arch package manager is version 4.3.14 and the one in the
nix-shell is version 4.3.12. At least I hope that is the problem because I
should be able to fix that, I hope.
On Thu, Aug 21, 2014 at 11:08 AM,
Wout,
That did the trick! Having some issues getting VBoxManage to work (either
inside the nix-shell or outside of it), but once that's fixed I'll be good
to go.
Thanks for all the help!
Rich
On Thu, Aug 21, 2014 at 3:13 AM, Wout Mertens
wrote:
> Hi Rich,
>
> you have to specify the attribut
On 21/08/2014 18:30, Alexander Zubkov wrote:
> > Being simple is for year '90
> It is hard to argue so strong opinion. :) By simple, I do not mean
> primitive. I mean using enough and appropriate tools to do the work,
> without extra unneeded stuff.
>
> Initrd is not a generic system in my mind.
On 08/21/2014 06:18 PM, Mateusz Kowalczyk wrote:
What's the procedure for dealing with such situations? At the moment
both packages in question are doCheck = false due to this. See [1] for
an example.
I would just point $HOME somewhere into the temporary build folder
(preferably only for the c
> Being simple is for year '90
It is hard to argue so strong opinion. :) By simple, I do not mean
primitive. I mean using enough and appropriate tools to do the work,
without extra unneeded stuff.
Initrd is not a generic system in my mind. It is just sequence of
actions to mount root.
1) If th
Hi,
There are some packages which require HOME to be writeable/available
during their build or test process.
A quick example are ‘cabal-bounds’ and ‘ghc-mod’ packages which both
need to call cabal-install during the testing phase but unfortunately
cabal insists on checking HOME/.cabal and causes
On 21/08/2014 17:35, Alexander Zubkov wrote:
> etc, etc, ... and viola - we will not be needing root filesystem at all. :)
>
> My option is keep initrd as simple as possible and not to bound it to
> particular init system. Because one day may be someone will make another
> option for init system
On 2014-08-21 17:55, Eelco Dolstra wrote:
> Hi,
>
> On 21/08/14 15:39, Luca Bruno wrote:
>
>> The advantages of using systemd are:
> [snip]
>
> Another advantage: starting journald very early.
>
> Also: running sshd in the initrd to support receiving the encryption key of
> the
> root disk via "ni
On 21/08/2014 15:55, Eelco Dolstra wrote:
>> I hope in an active discussion about this issue, and in a possible
>> resolution.
> I'm in favor in principle.
>
> The main technical issue is how to describe units for the initrd. I guess we
> can
> just have a ‘boot.initrd.systemd.*’ tree that mirrors
Hi,
On 21/08/14 15:39, Luca Bruno wrote:
> The advantages of using systemd are:
[snip]
Another advantage: starting journald very early.
Also: running sshd in the initrd to support receiving the encryption key of the
root disk via "nixops send-keys" :-)
> I hope in an active discussion about th
systemd is able to run in initrd, from the start to the end [1][2],
including the switch root:
My proposal is to use systemd to drive the initrd process for one reason
in particular: ordering of file system mounts. Two examples:
- My PR [3] about having nixos installed in a subdirectory, which
req
Hi Rich,
you have to specify the attribute for virtualbox, not the package name. So
linuxPackages.virtualbox in this case. You can find the attributes with
nix-env -qaP
Wout.
On Thu, Aug 21, 2014 at 12:02 AM, Richard Wallace <
rwall...@thewallacepack.net> wrote:
> Ok, that will work if I can
Hello all,
I'm an enthusiastic recent convert to nix. I'm trying to get it set up
across our university's shared RHEL 6.5 linux setup so that my students and
I can use it for building all our research software. I've convinced the
sysadmins to symlink /nix/ to an NSF location (for now) so that we
Whew... after all that, it looks like the problem was simply where
~/.nix-profile/ was pointing. It was pointing to
/nix/var/nix/profiles/default...
I did not know that nix-env update the profile pointed to by that symlink.
But apparently that's what it does.
On Thu, Aug 21, 2014 at 12:18 AM,
mit-scheme: fix for doc compilation
---
diff --git a/pkgs/development/compilers/mit-scheme/default.nix
b/pkgs/development/compilers/mit-scheme/default.nix
index c025fc8..1e0beb1 100644
--- a/pkgs/development/compilers/mit-scheme/default.nix
+++ b/pkgs/development/compilers/mit-scheme/default.nix
> I don't have such a list but you can grep for ‘broken = true;’ in
> pkgs/development/{libraries,tools}/haskell/*/*.nix. Often the commit
> marking it as broken will refer to open upstream issue.
>
> Yes, there is work to be done which is to now try to get the broken
> packages into a non-broken s
On 08/21/2014 07:14 AM, Raahul Kumar wrote:
> Thanks! So can I have a list of broken package sorted by Hackage
> dependencies, so I know which packages are used the most?
> Sounds like there is still work to be done.
>
> Aloha,
>
> RK.
>
>
> On Thu, Aug 21, 2014 at 3:19 PM, Mateusz Kowalczyk
>
22 matches
Mail list logo
