question about fstab in squeeze and uuid
in [Debian]
|
From: Camaleón on 12 Mar 2010 15:20 On Fri, 12 Mar 2010 20:21:14 +0100, Javier Barroso wrote: (...) >> Maybe there is a good comparison chart about all these methods that >> list their "pros" and "cons" :-? > Not a chart, but yes references to why uuid ... : > https://bugzilla.redhat.com/show_bug.cgi?id=364441 > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=572376 After a bit of digging, it seems all methods have its own drawbacks: by-label -> can generate collissions and is not present in all volumes by-path -> can also change because it depends on the system bus it is attached to by-id -> it seems that not all devices provide an ID by-uuid -> as we have seen, when formating or repartitioning the device it can also change Well, that said I like Lenny still uses the old scheme "/dev/sdx". At least if it changes, I still understand it better than the new udev naming :-) Greetings, -- Camaleón -- To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org Archive: http://lists.debian.org/pan.2010.03.12.20.16.48(a)gmail.com
From: Cousin Stanley on 12 Mar 2010 15:40 > .... > I personally don't like the use of UUIDS by the installer > and I change them back to their traditional names. > .... I'm also not a uuid fan and find that using disk labels in /etc/fstab a la /dev/disk/by-label as an alternative has worked well for me .... -- Stanley C. Kitching Human Being Phoenix, Arizona -- To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org Archive: http://lists.debian.org/hne7ea$8dd$1(a)news.eternal-september.org
From: thib on 12 Mar 2010 20:10 Stephen Powell wrote: > I believe a UUID is generated when the partition is "formatted", either with > mkfs or mkswap. I confirm - just tried shrinking and growing back an extfs. UUID is left untouched (as expected); that Mint article is BS or just obsolete. -thib -- To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org Archive: http://lists.debian.org/4B9AE392.6000909(a)stammed.net
From: Paul E Condon on 13 Mar 2010 00:30 On 20100313_020002, thib wrote: > Stephen Powell wrote: > >I believe a UUID is generated when the partition is "formatted", either with > >mkfs or mkswap. > > I confirm - just tried shrinking and growing back an extfs. UUID is > left untouched (as expected); that Mint article is BS or just > obsolete. > > -thib A bit worrisome to me. UUID must be persistent during normal life of a device, so it can be used as an identifier. A copy of the UUID that is readable must be attached to the device, so that the device can be identified and distinguished from any other device. Either an arbitrary random number can be recorded on read-mostly-memory in the device, or some fixed data in the format of the device is used in conjunction with a message-digest algorithm to come up with a 32 hex digit id-number. And only fixed data on the device can be used. If the resulting 32hex digit number is not, itself, recorded in some hard to find place on the device. Your experiment shows that UUID is dependent on partition table data. Without bothering to check, I assert that it is not dependent on any other non-partition data that is saved on the device. If it were it wouldn't be persistent between reboots, etc. The number of bytes in a partition table is quite small and those bytes conform to a rather structured pattern, so there are far fewer independent patterns than 128 independent bits of partition record. In fact I would expect that name collisions are actually far more likely than one would expect from a simple-minded probability calculation. -- Paul E Condon pecondon(a)mesanetworks.net -- To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org Archive: http://lists.debian.org/20100313052658.GA2170(a)big.lan.gnu
From: Freeman on 13 Mar 2010 04:00 On Fri, Mar 12, 2010 at 08:16:48PM +0000, Camale�n wrote: > On Fri, 12 Mar 2010 20:21:14 +0100, Javier Barroso wrote: > > (...) > > >> Maybe there is a good comparison chart about all these methods that > >> list their "pros" and "cons" :-? > > > Not a chart, but yes references to why uuid ... : > > https://bugzilla.redhat.com/show_bug.cgi?id=364441 > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=572376 > > After a bit of digging, it seems all methods have its own drawbacks: > > by-label -> can generate collissions and is not present in all volumes > > by-path -> can also change because it depends on the system bus it is > attached to > > by-id -> it seems that not all devices provide an ID > > by-uuid -> as we have seen, when formating or repartitioning the device > it can also change > > Well, that said I like Lenny still uses the old scheme "/dev/sdx". At > least if it changes, I still understand it better than the new udev > naming :-) > I've been changing to labels with squeeze. The up side: it is a hard designation and can be made unique. The collisions part is unclear to me. If one plugs into USB a drive from another machine, won't that just be listed in /dev by sdx, at which point it can have its fstab edited? I typo-ed the label for my root partition on my last fstab update but it mounted anyway as rootfs in mtab. So I put rootfs in fstab and it has been working. :-/ -- Kind Regards, Freeman http://bugs.debian.org/release-critical/ -- To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org Archive: http://lists.debian.org/20100313085335.GA5439(a)Europa.office
First
|
Prev
|
Next
|
Last
Pages: 1 2 3 4 5 Prev: pen drive without format and can not find it Next: Incredibly useful Firefox addon: Hyperwords |