Re: freebsd-hackers Digest, Vol 624, Issue 6
"Pokala, Ravi" <[email protected]>
| Newsgroups | gmane.os.freebsd.devel.hardware,gmane.os.freebsd.devel.hackers |
|---|---|
| Message-ID | <D1487BCD.13254D%[email protected]> |
>Date: Sat, 4 Apr 2015 02:28:46 +0000 >From: "Pokala, Ravi" <[email protected]> >To: "[email protected]" <[email protected]>, > "[email protected]" <[email protected]> >Subject: Booting off NVMe using traditional bootstrap? >Message-ID: <D1449A6E.1322B6%[email protected]> >Content-Type: text/plain; charset="us-ascii" > >Hi folks, > >Does anyone know off-hand if it's possible to boot (amd64) off of an NVMe >device using the traditional bootstrap code (i.e. *not* UEFI)? > >Thanks, > >Ravi Naturally, someone pointed out the obvious idea - put /boot on a USB stick and boot off that. As silly as it sounds, we've had trouble doing that in the past - USB hiccups caused '/' to disappear, and hilarity ensued. Nothing like an expensive server going offline because a $10 component failed. :-P Though I suppose in the past we were putting all of '/' on the USB, not just '/boot'... So, how little can we get away with putting on the USB? For it to have '/boot', doesn't that mean it also has have '/'? Or else, how would it recognize that the 'boot' directory on the USB was actually '/boot'? Doing this (minimal bootstrap on one device, everything else on other devices) seems like it should be documented, but a quick trip to Google and to freebsd.org don't reveal anything obvious. Any pointers? Thanks, Ravi _______________________________________________ [email protected] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hardware To unsubscribe, send any mail to "[email protected]"