unofficial mirror of help-guix@gnu.org 
 help / color / mirror / Atom feed
From: Tobias Geerinckx-Rice <me@tobias.gr>
To: help-guix@gnu.org
Subject: Re: Installer: GUIX_IMAGE as /dev/sda on some hardware?
Date: Sat, 25 May 2019 13:58:20 +0200	[thread overview]
Message-ID: <87r28mzqxv.fsf@nckx> (raw)
In-Reply-To: <874l5i52i1.fsf@roquette.mug.biscuolo.net>

[-- Attachment #1: Type: text/plain, Size: 1787 bytes --]

Giovanni Biscuolo wrote:
> This is **very** important when installing grub, in fact grub
> installation failed when instantiating my config.scm on the HP 
> ProLiant
> simply because it was on /dev/sda pointing to the USB media;

/dev/xdyN names have never been safe to use in this way, though. 
I guess they're fine (at your own risk…) as short identifiers 
during a single CLI session, but not across reboots.  It just 
often *happens* to work on most machines, which as we all know, is 
the most dangerous kind of working.

> 3. how is the USB media "relocated" to the last /dev/sd? device 
> by the
> installer?

It's… not?  Dev nodes & names are doled out by the kernel.  As 
you've discovered, they aren't to be relied on, and you should use 
labels or UUIDs instead.

Even if we'd pretend differently in the installer, for example 
through artificially delayed probing for usb_storage devices, 
things would still break when that lie would be exposed by adding 
a new SATA drive, or when using a different (rescue) distro, or…

> 1. has anyone observed a similar issue?
>
> 2. what could have caused it?

I used to own a system that scanned drives in a different order 
depending on whether or not a USB keyboard was attached.  It was 
intended as a headless system; great fun debugging that!  

I once corrupted a drive (luckily part of a RAID-1 array, but it 
was still stupid of me) because the rescue CD scanned drives in a 
different order than the distro I was used to.

My first SATA motherboard enumerated drives differently between 
cold and warm boots — but only sometimes.  I think it had 
something to do with spin-up times.

I could go on.  I often do.  Sorry.

Hoping to have scared you into using UUIDs,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 227 bytes --]

  reply	other threads:[~2019-05-25 12:14 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-25  9:04 Installer: GUIX_IMAGE as /dev/sda on some hardware? Giovanni Biscuolo
2019-05-25 11:58 ` Tobias Geerinckx-Rice [this message]
2019-05-25 13:16   ` Giovanni Biscuolo
2019-05-26 11:22     ` Tobias Geerinckx-Rice
2019-05-26 11:31     ` Tobias Geerinckx-Rice
2019-05-28 15:11     ` Ludovic Courtès

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://guix.gnu.org/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87r28mzqxv.fsf@nckx \
    --to=me@tobias.gr \
    --cc=help-guix@gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).