[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Inscrutable error when using “guix deploy”
From: |
Jakob L. Kreuze |
Subject: |
Re: Inscrutable error when using “guix deploy” |
Date: |
Sat, 24 Aug 2019 13:55:54 -0400 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/26.2 (gnu/linux) |
Hi Ricardo and Alex,
Ricardo Wurmus <address@hidden> writes:
>> This error, in general shape, looks very much like an error I kept getting
>> when first "deploying" to a guix server. I found that doing a pull on the
>> server, followed by a system reconfigure then allowed me to actually
>> successfully run deploy afterwards.
>>
>> I'm not sure what causes that, but thought perhaps it was to do with
>> running an older version of guix on the target server before pulling?
>
> Thanks, it’s possible that this is in fact the case here.
>
> The error is really hard to diagnose, though, because there’s no file to
> inspect. It would be good if “guix deploy” could catch errors stemming
> from the use of older Guixes on the target and throw a more instructive
> error message.
I saw this specific error message a few times while I was building 'guix
deploy'. It means something's been printed that can't be read, which
(might) be the case if something doesn't have a G-Expression compiler.
My money is on the target not having a reader for <operating-system>,
but as you've observed, the error reporting makes this incredibly
difficult to diagnose.
> @Jakob: What do you think? Can we check for required Guix facilities on
> the remote and abort if they don’t exist?
I think that's a good idea, especially since we've had issues with
deploying to older Guixes before (with the 'user-homes' service).
Regards,
Jakob
signature.asc
Description: PGP signature