[Bug 565646] New: mount fails silently to mount a partition.
http://bugzilla.novell.com/show_bug.cgi?id=565646 http://bugzilla.novell.com/show_bug.cgi?id=565646#c0 Summary: mount fails silently to mount a partition. Classification: openSUSE Product: openSUSE 11.2 Version: Final Platform: Other OS/Version: Other Status: NEW Severity: Normal Priority: P5 - None Component: Basesystem AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: carlos.e.r@opensuse.org QAContact: qa@suse.de Found By: --- Blocker: --- fstab: LABEL=test_6 /mnt/TestBig xfs \ noauto,user,noatime,nodiratime,exec,nofail 1 4 bombadillo:/mnt # mount /mnt/TestBig <============ fails silently bombadillo:/mnt # umount /mnt/TestBig umount: /mnt/TestBig: not mounted The mount command fails silently, because the label is wrong (my mistake): bombadillo:/mnt # l /dev/disk/by-label/ | grep sdb lrwxrwxrwx 1 root root 10 Dec 17 17:11 test_5 -> ../../sdb5 lrwxrwxrwx 1 root root 10 Dec 17 17:13 test_big -> ../../sdb6 I understood that the linux rationale is that programs succeed silently, but give a message on errors. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c
shuang qiu
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c1
Petr Uzel
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c3
Petr Uzel
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c4
Carlos Robinson
Try to remove the 'nofail' option from the fstab entry.
Well, yes, removing "nofail" avoids this problem, but causes another one, much worse: Boot sequence fails and stops if that external disk is not present. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c5
Petr Uzel
Boot sequence fails and stops if that external disk is not present.
Sorry, but I don't understand what would you expect to happen if you have invalid entry in /etc/fstab? IMHO you should use 'noauto' option, so that the external drive does not get mounted automatically, but only upon explicit request. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c6
Carlos Robinson
(In reply to comment #4)
Boot sequence fails and stops if that external disk is not present.
Sorry, but I don't understand what would you expect to happen if you have invalid entry in /etc/fstab?
IMHO you should use 'noauto' option, so that the external drive does not get mounted automatically, but only upon explicit request.
I have to use both options. With "noauto" it doesn't attempt to mount during boot, true, but it will try to fsck it; if the device does not exist, the boot sequence fails completely dropping into emergency rescue mode. The solution, which was given to me on Bug #461294 and Bug #375248, is to add the "nofail" option. The idea is simply to get a successful boot sequence. The unexpected consequence is what I reported on this bugzilla, that on a manual attempt it doesn't even give a warning message (not an error). -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c7
Petr Uzel
I have to use both options.
Sorry, I did not notice that you're using 'noauto' already.
With "noauto" it doesn't attempt to mount during boot, true, but it will try to fsck it;
What if you set fs_passno to 0 (6th filed in /etc/fstab) - does it still fail on fsck if the device is not present?
if the device does not exist, the boot sequence fails completely dropping into emergency rescue mode. The solution, which was given to me on Bug #461294 and Bug #375248, is to add the "nofail" option.
The idea is simply to get a successful boot sequence. The unexpected consequence is what I reported on this bugzilla, that on a manual attempt it doesn't even give a warning message (not an error).
Well, that's exactly the documented and expected behavior of 'nofail' option. If you want this to be changed (to give a warning), then it is something for upstream IMHO. What I think is most reasonable (to get successful boot sequence): - fix the LABEL in /etc/fstab entry - add nofail back (and keep noauto) options, so that the boot sequence does not fail if the device is not present at boot time -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c8
Carlos Robinson
(In reply to comment #6)
What if you set fs_passno to 0 (6th filed in /etc/fstab) - does it still fail on fsck if the device is not present?
I'll have to test that (I don't boot that often). I don't remember if I tested that already... Ok, yes, I see I have entries that way in another computer: label=... noauto, 0 0. A manual mount reports that the device doesn't exist, which is good. The problem is, I suppose, that those disks will never be checked automatically. Can't have everything, I guess. Ah, look, another reason why - I have this email copy pasted in my fstab file as a comment: +++········ http://lists.opensuse.org/opensuse-factory/2009-10/msg00524.html Re: [opensuse-factory] Mounting existing crypted partition fails on 11.2rc1 .. Yes, that fsck behavior isn't optimal. Matthias König tried to improve it¹ and the outcome was that /etc/init.d/boot.localfs skips devices with mount option 'nofail' entirely. boot.crypto will take care of fsck then. That doesn't work with devices that need to by unlocked by boot.crypto-early already though :-( So for those you better set timeout=0,tries=0 in crypttab. I guess I should document that in the manpage. ········++- So, the best option to use in the fstab file is "nofail". But sometimes it isn't... life is complicated. Sigh.
if the device does not exist, the boot sequence fails completely dropping into emergency rescue mode. The solution, which was given to me on Bug #461294 and Bug #375248, is to add the "nofail" option.
The idea is simply to get a successful boot sequence. The unexpected consequence is what I reported on this bugzilla, that on a manual attempt it doesn't even give a warning message (not an error).
Well, that's exactly the documented and expected behavior of 'nofail' option. If you want this to be changed (to give a warning), then it is something for upstream IMHO.
I understand. Now. Thanks :-) -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=565646
http://bugzilla.novell.com/show_bug.cgi?id=565646#c9
Petr Uzel
What if you set fs_passno to 0 (6th filed in /etc/fstab) - does it still fail on fsck if the device is not present?
I'll have to test that (I don't boot that often). I don't remember if I tested that already... Ok, yes, I see I have entries that way in another computer: label=... noauto, 0 0. A manual mount reports that the device doesn't exist, which is good.
If you manage to test it, please post the results here for reference.
The idea is simply to get a successful boot sequence. The unexpected consequence is what I reported on this bugzilla, that on a manual attempt it doesn't even give a warning message (not an error).
Well, that's exactly the documented and expected behavior of 'nofail' option. If you want this to be changed (to give a warning), then it is something for upstream IMHO.
I understand. Now. Thanks :-)
If you want to have mount(8) more verbose if the device does not exist and 'nofail' option is given, I suggest to discuss the change first on upstream ML (util-linux-ng@vger.kernel.org). Thanks. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply@novell.com