Date: Fri, 2 Mar 2018 16:57:00 -0800 From: bob prohaska <fbsd@www.zefox.net> To: tech-lists <tech-lists@zyxst.net> Cc: freebsd-arm@freebsd.org, bob prohaska <fbsd@www.zefox.net> Subject: Re: Can two USB flash drives conflict with each other? Message-ID: <20180303005700.GC37148@www.zefox.net> In-Reply-To: <0fd4b991-a7d8-0e04-7d73-26d351873390@zyxst.net> References: <20180302231317.GA37148@www.zefox.net> <0fd4b991-a7d8-0e04-7d73-26d351873390@zyxst.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Mar 03, 2018 at 12:32:27AM +0000, tech-lists wrote: > On 02/03/2018 23:13, bob prohaska wrote: > > The obvious solution is "don't do that!", but if somebody can offer > > a more insightful explanation I'd be grateful. Using two USB flash > > drives simultaneously would be very useful. > > I've found [this was a year ago, maybe two] that if I had two usb sticks > plugged in that sometimes they'd be detected in reverse order to what I > expected. > > What I mean is that sometimes the device called /dev/da0 and the one > called /dev/da1 would swap on reboot. I suppose it would depend on which > one woke up first. So if I had made /dev/da0p1, allocated it as swap, > /dev/da0p1 as data, perhaps put the ports tree there, /dev/da1p1 as > data, perhaps used the entire device for data, sometimes it'd boot, look > at /dev/da0 which was /dev/da1 previously, not seen swap, and complained > loudly. > > I think there is a way to wire device identities to names but it might > need GPT rather than MBR as a partitioning scheme. I worked around it by > labelling one of the usb sticks with sticky tape and ensuring it wasn't > plugged in before the other one when rebooting. > On the first try I plugged the second USB drive into a running machine, producing the errors reported. It's not obvious how a _second_ device can "unseat" one that is already represented in /dev/.... On a later try I plugged the second USB flash device in and powered the Pi3 up, whence the kernel got confused over which was which. That makes slightly more sense. I think that might be fixable with labels in /etc/fstab. In my case the second drive was labeled much like the first, so it couldn't help. Somewhere I got the idea USB flash devices had a unique serial number, or equivalent, so that more than one could co-exist on a host. Is this notion mistaken? Thanks for reading, bob prohaska
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20180303005700.GC37148>