[opensuse] Obsolete entry in fstab stalls booting
Please note, that the problem at hand is already solved, but the underlying issue gets in my way from time to time, so I wanted to tell (or rant 😉) about it. Two days ago, I wasn't able to boot my system (LEAP 15.1): X-Server didn't start and I could only log into maintenance mode with no network either. First, I suspected a botched update, but the kernel-firmware was updated a couple of days before and I had successfully booted in the mean time. So, I was dumbfounded, but then I remembered to look into /etc/fstab and found an entry of a USB drive that I had plugged in and removed. I deleted said entry and LEAP booted just fine. My point is: maybe the maintainers could help to avoid this situation: Please, don't break the boot process if some malfunctioning but unimportant entry in fstab doesn't work! Or at least, make it easier for the user to find out what is going wrong. I had similar issues multiple times over the past years and sometimes it took me quite a time to figure them out. Regards. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
zb4ng wrote:
Please note, that the problem at hand is already solved, but the underlying issue gets in my way from time to time, so I wanted to tell (or rant 😉) about it.
Two days ago, I wasn't able to boot my system (LEAP 15.1): X-Server didn't start and I could only log into maintenance mode with no network either. First, I suspected a botched update, but the kernel-firmware was updated a couple of days before and I had successfully booted in the mean time. So, I was dumbfounded, but then I remembered to look into /etc/fstab and found an entry of a USB drive that I had plugged in and removed. I deleted said entry and LEAP booted just fine.
There is no reason for a USB stick to end up in fstab - unless maybe if you booted from USB ? -- Per Jessen, Zürich (6.9°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 21.02.20 um 11:23 schrieb Per Jessen:
zb4ng wrote:
Please note, that the problem at hand is already solved, but the underlying issue gets in my way from time to time, so I wanted to tell (or rant 😉) about it.
Two days ago, I wasn't able to boot my system (LEAP 15.1): X-Server didn't start and I could only log into maintenance mode with no network either. First, I suspected a botched update, but the kernel-firmware was updated a couple of days before and I had successfully booted in the mean time. So, I was dumbfounded, but then I remembered to look into /etc/fstab and found an entry of a USB drive that I had plugged in and removed. I deleted said entry and LEAP booted just fine.
There is no reason for a USB stick to end up in fstab - unless maybe if you booted from USB ?
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it. But hours later, when I rebooted, I just got the "maintenance" prompt and couldn't see the connection! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 11.49, zb4ng wrote:
Am 21.02.20 um 11:23 schrieb Per Jessen:
zb4ng wrote:
Please note, that the problem at hand is already solved, but the underlying issue gets in my way from time to time, so I wanted to tell (or rant 😉) about it.
Two days ago, I wasn't able to boot my system (LEAP 15.1): X-Server didn't start and I could only log into maintenance mode with no network either. First, I suspected a botched update, but the kernel-firmware was updated a couple of days before and I had successfully booted in the mean time. So, I was dumbfounded, but then I remembered to look into /etc/fstab and found an entry of a USB drive that I had plugged in and removed. I deleted said entry and LEAP booted just fine.
There is no reason for a USB stick to end up in fstab - unless maybe if you booted from USB ?
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it.
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem. Next time, install and use "gparted" instead.
But hours later, when I rebooted, I just got the "maintenance" prompt and couldn't see the connection!
Well, the emergency prompt should have told you the reason: I can not mount "whatever". It may have said so but flowed out of the screen. I find the emergency text since systemd not helpful. I would suggest that you do it again and look carefully at the text, then if the reason is not clear, take a photo and report in bugzilla. Unfortunately, the boot process has no way of knowing if some disk mentioned in fstab is important or not, it can only abort boot and ask. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
On 2020-02-21 09:00 AM, Carlos E. R. wrote:
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem.
Unless I'm mistaken, that line is not created unless you specifically cause it to be created. To do that, you have to click on the "Mount device" button. Do not mount is the default. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 21/02/2020 à 15:00, Carlos E. R. a écrit :
On 21/02/2020 11.49, zb4ng wrote:
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it.
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem.
I don't think it's default, but yast ask "do you want it mounted" but don't add "permanently" - may be an enhancement possible :-)
Next time, install and use "gparted" instead.
frankly, no!! if you want command line, fdisk and mkfs do a superb job
Well, the emergency prompt should have told you the reason: I can not mount "whatever".
yes, but AFAIK disk by UUID, very cryptic. I experienced this several times (but with swap being reformatted by an other install, UUID change) jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
jdd@dodin.org wrote:
Le 21/02/2020 à 15:00, Carlos E. R. a écrit :
On 21/02/2020 11.49, zb4ng wrote:
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it.
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem.
I don't think it's default, but yast ask "do you want it mounted" but don't add "permanently" - may be an enhancement possible :-)
I'd even call that a bug. As a user, if I format a USB stick, I want to use it, and for that I need to mount it. So it is straightforward to check this option. It's therefore (IMHO) YaSTs task to make sure nothing bad happens. Just adding such a line without making the user aware of the consequences is quite careless. Such an entry may be usefull for fixed disks, but for removable devices that are anyhow handled separately from fstab this should (again, IMO) not be done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 16.25, Peter Suetterlin wrote:
jdd@dodin.org wrote:
Le 21/02/2020 à 15:00, Carlos E. R. a écrit :
On 21/02/2020 11.49, zb4ng wrote:
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it.
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem.
I don't think it's default, but yast ask "do you want it mounted" but don't add "permanently" - may be an enhancement possible :-)
I'd even call that a bug. As a user, if I format a USB stick, I want to use it, and for that I need to mount it. So it is straightforward to check this option. It's therefore (IMHO) YaSTs task to make sure nothing bad happens. Just adding such a line without making the user aware of the consequences is quite careless.
if you tell YaST "I want to mount it" you are telling it "I want to mount it by using an fstab entry, so create it for me, please". It has been this way for decades.
Such an entry may be usefull for fixed disks, but for removable devices that are anyhow handled separately from fstab this should (again, IMO) not be done.
It is not that easy to know it is a removable drive. Sometimes it works: Telcontar:~ # lsblk --help | grep -i removable RM removable device HOTPLUG removable or hotplug device (usb, pcmcia, ...) Telcontar:~ # I put a stick: Telcontar:~ # lsblk --output NAME,KNAME,SIZE,TYPE,RM,HOTPLUG,LABEL /dev/sdf NAME KNAME SIZE TYPE RM HOTPLUG LABEL sdf sdf 7.2G disk 1 1 openSUSE_Leap_15.0_Rescue_CD ├─sdf1 sdf1 613M part 1 1 openSUSE_Leap_15.0_Rescue_CD ├─sdf2 sdf2 15M part 1 1 BOOT └─sdf3 sdf3 6.6G part 1 1 cow Telcontar:~ # Some people use removable drives as permanent. I do. Isengard:~ # lsblk --output NAME,KNAME,SIZE,TYPE,RM,HOTPLUG,LABEL NAME KNAME SIZE TYPE RM HOTPLUG LABEL ... sdd sdd 7.3T disk 0 1 └─sdd1 sdd1 7.3T part 0 1 Telcontar_1 └─cr_my_book_tlcntr dm-4 7.3T crypt 0 0 ANameUnseen That machine has 4 external hard disks via usb3, permanently mounted. You are the root, the admin, you have a job :-D -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
On Fri, 21 Feb 2020 19:37:27 +0100 "Carlos E. R." <robin.listas@telefonica.net> wrote:
On 21/02/2020 16.25, Peter Suetterlin wrote:
jdd@dodin.org wrote:
Le 21/02/2020 à 15:00, Carlos E. R. a écrit :
On 21/02/2020 11.49, zb4ng wrote:
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it.
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem.
I don't think it's default, but yast ask "do you want it mounted" but don't add "permanently" - may be an enhancement possible :-)
I'd even call that a bug. As a user, if I format a USB stick, I want to use it, and for that I need to mount it. So it is straightforward to check this option. It's therefore (IMHO) YaSTs task to make sure nothing bad happens. Just adding such a line without making the user aware of the consequences is quite careless.
if you tell YaST "I want to mount it" you are telling it "I want to mount it by using an fstab entry, so create it for me, please". It has been this way for decades.
That may be the case, I don't know, but if so then it should make it clear. I wouldn't have expected it for one. I would expect that if something asks me whether I want it to mount something then it is asking me whether I want it to execute a mount command on my behalf, not change a file!
Such an entry may be usefull for fixed disks, but for removable devices that are anyhow handled separately from fstab this should (again, IMO) not be done.
It is not that easy to know it is a removable drive. Sometimes it works:
Telcontar:~ # lsblk --help | grep -i removable RM removable device HOTPLUG removable or hotplug device (usb, pcmcia, ...) Telcontar:~ #
I put a stick:
Telcontar:~ # lsblk --output NAME,KNAME,SIZE,TYPE,RM,HOTPLUG,LABEL /dev/sdf NAME KNAME SIZE TYPE RM HOTPLUG LABEL sdf sdf 7.2G disk 1 1 openSUSE_Leap_15.0_Rescue_CD ├─sdf1 sdf1 613M part 1 1 openSUSE_Leap_15.0_Rescue_CD ├─sdf2 sdf2 15M part 1 1 BOOT └─sdf3 sdf3 6.6G part 1 1 cow Telcontar:~ #
Some people use removable drives as permanent. I do.
Isengard:~ # lsblk --output NAME,KNAME,SIZE,TYPE,RM,HOTPLUG,LABEL NAME KNAME SIZE TYPE RM HOTPLUG LABEL ... sdd sdd 7.3T disk 0 1 └─sdd1 sdd1 7.3T part 0 1 Telcontar_1 └─cr_my_book_tlcntr dm-4 7.3T crypt 0 0 ANameUnseen
That machine has 4 external hard disks via usb3, permanently mounted.
You are the root, the admin, you have a job :-D
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Dave Howorth wrote:
On Fri, 21 Feb 2020 19:37:27 +0100 "Carlos E. R." <robin.listas@telefonica.net> wrote:
if you tell YaST "I want to mount it" you are telling it "I want to mount it by using an fstab entry, so create it for me, please". It has been this way for decades.
That may be the case, I don't know, but if so then it should make it clear. I wouldn't have expected it for one. I would expect that if something asks me whether I want it to mount something then it is asking me whether I want it to execute a mount command on my behalf, not change a file!
Perhaps(!) take the name as a hint - YaST = Yet another Setup Tool. It really is a configuration management tool. The partitioner helps you manage fstab and your filesystems, especially at installation time. I would expect whatever changes I make with YaST to be permanent. -- Per Jessen, Zürich (3.3°C) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
if you tell YaST "I want to mount it" you are telling it "I want to mount it by using an fstab entry, so create it for me, please". It has been this way for decades.
I just had to use yast partitionner, it would be pretty easy to add only a text "mount with fstab" on the yast windows. No need to change the code... jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 20.47, Dave Howorth wrote:
On Fri, 21 Feb 2020 19:37:27 +0100 "Carlos E. R." <robin.listas@telefonica.net> wrote:
On 21/02/2020 16.25, Peter Suetterlin wrote:
jdd@dodin.org wrote:
Le 21/02/2020 à 15:00, Carlos E. R. a écrit :
On 21/02/2020 11.49, zb4ng wrote:
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it.
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem.
I don't think it's default, but yast ask "do you want it mounted" but don't add "permanently" - may be an enhancement possible :-)
I'd even call that a bug. As a user, if I format a USB stick, I want to use it, and for that I need to mount it. So it is straightforward to check this option. It's therefore (IMHO) YaSTs task to make sure nothing bad happens. Just adding such a line without making the user aware of the consequences is quite careless.
if you tell YaST "I want to mount it" you are telling it "I want to mount it by using an fstab entry, so create it for me, please". It has been this way for decades.
That may be the case, I don't know, but if so then it should make it clear. I wouldn't have expected it for one. I would expect that if something asks me whether I want it to mount something then it is asking me whether I want it to execute a mount command on my behalf, not change a file!
It is not something, it is YaST. It is a system configuration tool, not a desktop tool for the user. Users have been told for years and years to use yast partitioner in order to add partitions to fstab. It is well known this behaviour. It is documented: <https://doc.opensuse.org/documentation/leap/reference/html/book.opensuse.reference/cha-expert-partitioner.html> And when you tick "mount device" it clearly shows "fstab options". -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
On 22/02/2020 05:21, Carlos E. R. wrote:
It is not something, it is YaST. It is a system configuration tool, not a desktop tool for the user.
Users have been told for years and years to use yast partitioner in order to add partitions to fstab. It is well known this behaviour.
It is documented: <https://doc.opensuse.org/documentation/leap/reference/html/book.opensuse.reference/cha-expert-partitioner.html>
And when you tick "mount device" it clearly shows "fstab options".
It handles LVM as well. What it doesn't handle, it seems to me, is dynamically resizing partitions with live file systems on them, or moving them around, shuffling to make a span of space where you've deleted smaller partitions on a larger drive. Yes, there are 3rd party tools for that :-) BUT this is why I use LVM and ReiserFS. It is the ultimate in 'deferred design' as far as file systems and disk partitioning goes. You can grow or shrink a ReiserFS system on LVM with the system running :-) No worries :-) BTDT. Of course I did it at the CLI 'cos YaST doesn't know about this. And yes it is well documented. I first learnt about this from a paper published by one of the SUSE technical staff, Michael Hasenstein, in 2001 that describes its implementation for SuSE 6.1 http://www.idevelopment.info/data/Unix/Linux/LINUX_lvm_whitepaper_SuSE.pdf A wonderful explanation! ftp://biologs.bf.lu.lv/pub/OS/datori/Compaq/Proliant350/linux/lvm_whitepaper.pdf -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 15.44, jdd@dodin.org wrote:
Le 21/02/2020 à 15:00, Carlos E. R. a écrit :
On 21/02/2020 11.49, zb4ng wrote:
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it.
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem.
I don't think it's default, but yast ask "do you want it mounted" but don't add "permanently" - may be an enhancement possible :-)
Next time, install and use "gparted" instead.
frankly, no!!
if you want command line, fdisk and mkfs do a superb job
gparted is a very nice GUI :-)
Well, the emergency prompt should have told you the reason: I can not mount "whatever".
yes, but AFAIK disk by UUID, very cryptic. I experienced this several times (but with swap being reformatted by an other install, UUID change)
Yes, UUID is "cryptic". It should also list label if present, maybe the bus. But at least prominently say "I expected a device to mount from fstab and I did not find it", then print the offending fstab line. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
Am 21.02.20 um 15:00 schrieb Carlos E. R.:
On 21/02/2020 11.49, zb4ng wrote:
Am 21.02.20 um 11:23 schrieb Per Jessen:
zb4ng wrote:
Please note, that the problem at hand is already solved, but the underlying issue gets in my way from time to time, so I wanted to tell (or rant 😉) about it.
Two days ago, I wasn't able to boot my system (LEAP 15.1): X-Server didn't start and I could only log into maintenance mode with no network either. First, I suspected a botched update, but the kernel-firmware was updated a couple of days before and I had successfully booted in the mean time. So, I was dumbfounded, but then I remembered to look into /etc/fstab and found an entry of a USB drive that I had plugged in and removed. I deleted said entry and LEAP booted just fine.
There is no reason for a USB stick to end up in fstab - unless maybe if you booted from USB ?
I just remembered that I tried to format the stick with YAST and then forgot about it and removed it.
That was your original problem, yast adds such a line. It is a convenience, but with usb sticks it causes a problem.
Next time, install and use "gparted" instead.
But hours later, when I rebooted, I just got the "maintenance" prompt and couldn't see the connection!
Well, the emergency prompt should have told you the reason: I can not mount "whatever". It may have said so but flowed out of the screen. I find the emergency text since systemd not helpful. I would suggest that
I found the message on the screen not helpful either, at least if you don't know what you are looking for. (The last message before the prompt was some NVidia-stuff). Also, I noticed that the network didn't work either, so this was misleading, too.
you do it again and look carefully at the text, then if the reason is not clear, take a photo and report in bugzilla.
I'll look into that.
Unfortunately, the boot process has no way of knowing if some disk mentioned in fstab is important or not, it can only abort boot and ask.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Unfortunately, the boot process has no way of knowing if some disk mentioned in fstab is important or not, it can only abort boot and ask.
but it could give a better warning, for example a fstab dump with the offending line in red (let alone to know the mount point) jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
zb4ng wrote:
I found the message on the screen not helpful either, at least if you don't know what you are looking for. (The last message before the prompt was some NVidia-stuff). Also, I noticed that the network didn't work either, so this was misleading, too.
That situation _is_ a bit difficult I agree. The pertinent information has most often scrolled past and you are left with trying to guess what happened. Typically when I can't deduce what has happened, this is when I hook up a serial console. -- Per Jessen, Zürich (5.0°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 18.51, Per Jessen wrote:
zb4ng wrote:
I found the message on the screen not helpful either, at least if you don't know what you are looking for. (The last message before the prompt was some NVidia-stuff). Also, I noticed that the network didn't work either, so this was misleading, too.
That situation _is_ a bit difficult I agree. The pertinent information has most often scrolled past and you are left with trying to guess what happened. Typically when I can't deduce what has happened, this is when I hook up a serial console.
The first thing I look at is fstab: I had this trouble many times. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
Carlos E. R. wrote:
On 21/02/2020 18.51, Per Jessen wrote:
zb4ng wrote:
I found the message on the screen not helpful either, at least if you don't know what you are looking for. (The last message before the prompt was some NVidia-stuff). Also, I noticed that the network didn't work either, so this was misleading, too.
That situation _is_ a bit difficult I agree. The pertinent information has most often scrolled past and you are left with trying to guess what happened. Typically when I can't deduce what has happened, this is when I hook up a serial console.
The first thing I look at is fstab: I had this trouble many times.
Things can get a little more complicated if you are e.g. booting from multi-pathed iscsi. or fibre. or NFS. -- Per Jessen, Zürich (3.4°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 16.09, zb4ng wrote:
Am 21.02.20 um 15:00 schrieb Carlos E. R.:
On 21/02/2020 11.49, zb4ng wrote:
Am 21.02.20 um 11:23 schrieb Per Jessen:
zb4ng wrote:
...
But hours later, when I rebooted, I just got the "maintenance" prompt and couldn't see the connection!
Well, the emergency prompt should have told you the reason: I can not mount "whatever". It may have said so but flowed out of the screen. I find the emergency text since systemd not helpful. I would suggest that
I found the message on the screen not helpful either, at least if you don't know what you are looking for. (The last message before the prompt was some NVidia-stuff). Also, I noticed that the network didn't work either, so this was misleading, too.
I guess those were problems the emergency boot was finding when booting in emergency mode, not the problems of the "normal" system :-(
you do it again and look carefully at the text, then if the reason is not clear, take a photo and report in bugzilla.
I'll look into that.
Thanks :-) -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
dealing with the symptom -- fail on boot -- is one issue; addressing the cause -- setting up your mounts correctly -- is another. the UI has a hard time being 'simple' for all use-cases; modifying direct config files can be useful. read-up on 'nofail' usage for systemd-managed mounts specified in fstab at, https://www.freedesktop.org/software/systemd/man/systemd.mount.html if you've not DISabled systemd's fstab use, systemd-fstab-generator converts fstab mount data to mount units. you can specify nofail/noauto/systemd.automount/etc options for a specific usb key in /etc/fstab a more generic approach, that matches any usb key can be implemented with systemd + udev rules. here's a good example https://www.andreafortuna.org/2019/06/26/automount-usb-devices-on-linux-usin... that^ can be modified to match on any particular mount device pattern. there _are_ more global options with manipulating systemd units -- setting DefaultDependencies=no, options overrides for specific targets ( e.g. local-fs.target ) -- but, proceed with caution on any global changes without really thinking thru want your intentions/actions are. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2020-02-21 05:23 AM, Per Jessen wrote:
There is no reason for a USB stick to end up in fstab - unless maybe if you booted from USB ?
I installed my 15.1 system from a USB stick and there is no mention of anything USB in my fstab. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
James Knott wrote:
On 2020-02-21 05:23 AM, Per Jessen wrote:
There is no reason for a USB stick to end up in fstab - unless maybe if you booted from USB ?
I installed my 15.1 system from a USB stick and there is no mention of anything USB in my fstab.
I was only trying to guess how it might have ended up in fstab. -- Per Jessen, Zürich (9.6°C) http://www.cloudsuisse.com/ - your owncloud, hosted in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2020-02-21 09:30 AM, Per Jessen wrote:
I installed my 15.1 system from a USB stick and there is no mention of anything USB in my fstab. I was only trying to guess how it might have ended up in fstab.
The only way I'm aware that it could happen is to deliberately do something to cause it. As I mentioned in another note, selecting "Mount device" when formatting in Yast would do it. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
James Knott wrote:
On 2020-02-21 09:30 AM, Per Jessen wrote:
I installed my 15.1 system from a USB stick and there is no mention of anything USB in my fstab. I was only trying to guess how it might have ended up in fstab.
The only way I'm aware that it could happen is to deliberately do something to cause it. As I mentioned in another note, selecting "Mount device" when formatting in Yast would do it.
Right, and I think we have also established that is exactly how it happened. -- Per Jessen, Zürich (9.6°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 15.19, James Knott wrote:
On 2020-02-21 05:23 AM, Per Jessen wrote:
There is no reason for a USB stick to end up in fstab - unless maybe if you booted from USB ?
I installed my 15.1 system from a USB stick and there is no mention of anything USB in my fstab.
You should get a repository from the stick. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
On 2020-02-21 09:43 AM, Carlos E. R. wrote:
On 21/02/2020 15.19, James Knott wrote:
On 2020-02-21 05:23 AM, Per Jessen wrote:
There is no reason for a USB stick to end up in fstab - unless maybe if you booted from USB ?
I installed my 15.1 system from a USB stick and there is no mention of anything USB in my fstab.
You should get a repository from the stick.
Quite so and back in the days when I installed from DVD, I'd also get it in there. But that is a repository, which has nothing to do with fstab. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
zb4ng wrote:
My point is: maybe the maintainers could help to avoid this situation: Please, don't break the boot process if some malfunctioning but unimportant entry in fstab doesn't work! Or at least, make it easier for the user to find out what is going wrong. I had similar issues multiple times over the past years and sometimes it took me quite a time to figure them out.
Stuff that is listed in /etc/fstab and is not explicitely marked as being non-crucial is *supposed* to block the boot process. If you put something in there, it's your responsibility to set proper parameters..... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 21.02.20 um 11:42 schrieb Peter Suetterlin:
zb4ng wrote:
My point is: maybe the maintainers could help to avoid this situation: Please, don't break the boot process if some malfunctioning but unimportant entry in fstab doesn't work! Or at least, make it easier for the user to find out what is going wrong. I had similar issues multiple times over the past years and sometimes it took me quite a time to figure them out.
Stuff that is listed in /etc/fstab and is not explicitely marked as being non-crucial is *supposed* to block the boot process. If you put something in there, it's your responsibility to set proper parameters.....
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial. Imho - assuming they want to get more users for openSuSE or Linux in general - getting into maintenance mode might scare them away, so changing the defaults in YAST etc might help making the system more beginner-friendly. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 13.54, zb4ng wrote:
Am 21.02.20 um 11:42 schrieb Peter Suetterlin:
zb4ng wrote:
My point is: maybe the maintainers could help to avoid this situation: Please, don't break the boot process if some malfunctioning but unimportant entry in fstab doesn't work! Or at least, make it easier for the user to find out what is going wrong. I had similar issues multiple times over the past years and sometimes it took me quite a time to figure them out.
Stuff that is listed in /etc/fstab and is not explicitely marked as being non-crucial is *supposed* to block the boot process. If you put something in there, it's your responsibility to set proper parameters.....
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
That assumption can not be made. There can be, for example, a data partition which if not mounted causes big multimedia files to write to "/" instead, fill it up, and crash the system. It is not theoretical, it is my case. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
On Fri, 21 Feb 2020 15:03:41 +0100 "Carlos E. R." <robin.listas@telefonica.net> wrote:
On 21/02/2020 13.54, zb4ng wrote:
Am 21.02.20 um 11:42 schrieb Peter Suetterlin:
zb4ng wrote:
My point is: maybe the maintainers could help to avoid this situation: Please, don't break the boot process if some malfunctioning but unimportant entry in fstab doesn't work! Or at least, make it easier for the user to find out what is going wrong. I had similar issues multiple times over the past years and sometimes it took me quite a time to figure them out.
Stuff that is listed in /etc/fstab and is not explicitely marked as being non-crucial is *supposed* to block the boot process. If you put something in there, it's your responsibility to set proper parameters.....
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
That assumption can not be made.
There can be, for example, a data partition which if not mounted causes big multimedia files to write to "/" instead, fill it up, and crash the system.
You mean because the multimedia files get written to the underlying mount point, which is in some filesystem such as / ? If so you can try the technique I always use when setting up a new mount point. Set the permissions of the underlying mount point to 000. That way nothing can write into the directory on the underlying filesystem. When the extra filesystem is mounted over the top, it's permissions are used instead.
It is not theoretical, it is my case.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 17.03, Dave Howorth wrote:
On Fri, 21 Feb 2020 15:03:41 +0100 "Carlos E. R." <> wrote:
On 21/02/2020 13.54, zb4ng wrote:
Am 21.02.20 um 11:42 schrieb Peter Suetterlin:
zb4ng wrote:
...
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
That assumption can not be made.
There can be, for example, a data partition which if not mounted causes big multimedia files to write to "/" instead, fill it up, and crash the system.
You mean because the multimedia files get written to the underlying mount point, which is in some filesystem such as / ?
Yes.
If so you can try the technique I always use when setting up a new mount point. Set the permissions of the underlying mount point to 000. That way nothing can write into the directory on the underlying filesystem. When the extra filesystem is mounted over the top, it's permissions are used instead.
Interesting trick :-) It would suffice to be read only, though. I typically create a file named "not_mounted" which I can use on scripts for testing. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
On Fri, 21 Feb 2020 20:05:54 +0100 "Carlos E. R." <robin.listas@telefonica.net> wrote:
On 21/02/2020 17.03, Dave Howorth wrote:
On Fri, 21 Feb 2020 15:03:41 +0100 "Carlos E. R." <> wrote:
On 21/02/2020 13.54, zb4ng wrote:
Am 21.02.20 um 11:42 schrieb Peter Suetterlin:
zb4ng wrote:
...
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
That assumption can not be made.
There can be, for example, a data partition which if not mounted causes big multimedia files to write to "/" instead, fill it up, and crash the system.
You mean because the multimedia files get written to the underlying mount point, which is in some filesystem such as / ?
Yes.
If so you can try the technique I always use when setting up a new mount point. Set the permissions of the underlying mount point to 000. That way nothing can write into the directory on the underlying filesystem. When the extra filesystem is mounted over the top, it's permissions are used instead.
Interesting trick :-)
It's not a trick, it's a design feauture of linux.
It would suffice to be read only, though. I typically create a file named "not_mounted" which I can use on scripts for testing.
Well yes, but 000 makes it absolutely clear what the intention is and has no downside that I'm aware of. And apparently your solution doesn't actually always solve the problem, since you're complaining! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/02/2020 20.51, Dave Howorth wrote:
On Fri, 21 Feb 2020 20:05:54 +0100 "Carlos E. R." <> wrote:
On 21/02/2020 17.03, Dave Howorth wrote:
On Fri, 21 Feb 2020 15:03:41 +0100 "Carlos E. R." <> wrote:
On 21/02/2020 13.54, zb4ng wrote:
Am 21.02.20 um 11:42 schrieb Peter Suetterlin:
zb4ng wrote:
...
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
That assumption can not be made.
There can be, for example, a data partition which if not mounted causes big multimedia files to write to "/" instead, fill it up, and crash the system.
You mean because the multimedia files get written to the underlying mount point, which is in some filesystem such as / ?
Yes.
If so you can try the technique I always use when setting up a new mount point. Set the permissions of the underlying mount point to 000. That way nothing can write into the directory on the underlying filesystem. When the extra filesystem is mounted over the top, it's permissions are used instead.
Interesting trick :-)
It's not a trick, it's a design feauture of linux.
Yeah, sure. Everything are features, more or less known. :-D I have seen the issue of writing unintended to the parent device mentioned in hundreds of sites, and none mentioned the permission trick, so I'm not alone, which qualifies it as a trick :-P
It would suffice to be read only, though. I typically create a file named "not_mounted" which I can use on scripts for testing.
Well yes, but 000 makes it absolutely clear what the intention
I never use numeric permissions.
is and has no downside that I'm aware of. And apparently your solution doesn't actually always solve the problem, since you're complaining!
No, I'm not complaining. Boot fails if the mount doesn't mount, so it works. And if the mount is optional, I check that it has mounted and the script tells that the mount is not there. Now I simply have another tool. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
zb4ng wrote:
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
My apologies. I had assumed you added it yourself - I wasn't aware YaST does this when you check the 'mount it' option. So not your fault.
Imho - assuming they want to get more users for openSuSE or Linux in general - getting into maintenance mode might scare them away, so changing the defaults in YAST etc might help making the system more beginner-friendly.
Yep, agree on that. But the solution is to fix YaST - the drop to maintainance mode on mount failure is definitely needed for many other cases. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 21.02.20 um 16:29 schrieb Peter Suetterlin:
zb4ng wrote:
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
My apologies. I had assumed you added it yourself - I wasn't aware YaST does this when you check the 'mount it' option. So not your fault.
Imho - assuming they want to get more users for openSuSE or Linux in general - getting into maintenance mode might scare them away, so changing the defaults in YAST etc might help making the system more beginner-friendly.
Yep, agree on that. But the solution is to fix YaST - the drop to maintainance mode on mount failure is definitely needed for many other cases.
If there are many cases where this is necessary, then I agree with that. But the same issue happens, if you have two Linux systems - say openSuSe + Ubuntu - on your hard disk and remove a partition in Ubuntu, then openSuSE doesn't know anything about it and it will go into maintenance mode while booting. In this case, I see the only possibility is to make the boot message about the missing partition more understandable. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
zb4ng wrote:
Am 21.02.20 um 16:29 schrieb Peter Suetterlin:
zb4ng wrote:
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
My apologies. I had assumed you added it yourself - I wasn't aware YaST does this when you check the 'mount it' option. So not your fault.
Imho - assuming they want to get more users for openSuSE or Linux in general - getting into maintenance mode might scare them away, so changing the defaults in YAST etc might help making the system more beginner-friendly.
Yep, agree on that. But the solution is to fix YaST - the drop to maintainance mode on mount failure is definitely needed for many other cases.
If there are many cases where this is necessary, then I agree with that.
One case is really enough. If a filesystem could not be mounted, you generally don't want the system just marching on.
But the same issue happens, if you have two Linux systems - say openSuSe + Ubuntu - on your hard disk and remove a partition in Ubuntu, then openSuSE doesn't know anything about it and it will go into maintenance mode while booting.
That is expected behaviour. If you have asked your openSUSE to mount that filesystem on that partition, and it suddenly does not exist, what would expect your system to do ? -- Per Jessen, Zürich (4.9°C) http://www.cloudsuisse.com/ - your owncloud, hosted in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Fri, 21 Feb 2020 18:44:19 +0100 Per Jessen <per@computer.org> wrote:
zb4ng wrote:
Am 21.02.20 um 16:29 schrieb Peter Suetterlin:
zb4ng wrote:
Well, I wasn't aware changing fstab, I just formatted the drive in YAST and I am learning just now that I could mark an entry as non-crucial. Moreover, looking at my own system, all partitions other than "/" or "/home" would be non-crucial.
My apologies. I had assumed you added it yourself - I wasn't aware YaST does this when you check the 'mount it' option. So not your fault.
Imho - assuming they want to get more users for openSuSE or Linux in general - getting into maintenance mode might scare them away, so changing the defaults in YAST etc might help making the system more beginner-friendly.
Yep, agree on that. But the solution is to fix YaST - the drop to maintainance mode on mount failure is definitely needed for many other cases.
If there are many cases where this is necessary, then I agree with that.
One case is really enough. If a filesystem could not be mounted, you generally don't want the system just marching on.
Err, see below.
But the same issue happens, if you have two Linux systems - say openSuSe + Ubuntu - on your hard disk and remove a partition in Ubuntu, then openSuSE doesn't know anything about it and it will go into maintenance mode while booting.
That is expected behaviour. If you have asked your openSUSE to mount that filesystem on that partition, and it suddenly does not exist, what would expect your system to do ?
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers. I have never used the "nofail" option, but it sounds (man systemd.mount) like that is what you need in such a situation. -- Per Jessen, Zürich (2.9°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
I have never used the "nofail" option, but it sounds (man systemd.mount) like that is what you need in such a situation.
"nofail" has caveats. For instance, if I issue command mount: Telcontar:~ # mount -v /mnt/Ext/Old_D Telcontar:~ # Yet it failed: Telcontar:~ # l /mnt/Ext/Old_D total 8 drwxr-xr-x 2 root root 4096 Feb 15 2015 ./ drwxrwxr-x 13 root root 4096 Apr 21 2018 ../ -rw-r--r-- 1 root root 0 Feb 15 2015 not_mounted Telcontar:~ # -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
Carlos E. R. wrote:
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
I have never used the "nofail" option, but it sounds (man systemd.mount) like that is what you need in such a situation.
"nofail" has caveats.
Well, we _are_ talking about linux :-)
For instance, if I issue command mount:
Telcontar:~ # mount -v /mnt/Ext/Old_D Telcontar:~ #
Yet it failed:
Telcontar:~ # l /mnt/Ext/Old_D total 8 drwxr-xr-x 2 root root 4096 Feb 15 2015 ./ drwxrwxr-x 13 root root 4096 Apr 21 2018 ../ -rw-r--r-- 1 root root 0 Feb 15 2015 not_mounted
It seems to me that is _exactly_ the behaviour what you asked for when you specified 'nofail' = do not report if device does not exist. -- Per Jessen, Zürich (11.6°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/02/2020 12.42, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
No, they don't. Thunderbird is typically configured the same way as Windows, which doesn't have an MTA. The client software simply talks directly with the MTA at the ISP. Even Alpine can do it. It is other things like cron which needs a local MTA, acting as a server process ;-)
I have never used the "nofail" option, but it sounds (man systemd.mount) like that is what you need in such a situation.
"nofail" has caveats.
Well, we _are_ talking about linux :-)
For instance, if I issue command mount:
Telcontar:~ # mount -v /mnt/Ext/Old_D Telcontar:~ #
Yet it failed:
Telcontar:~ # l /mnt/Ext/Old_D total 8 drwxr-xr-x 2 root root 4096 Feb 15 2015 ./ drwxrwxr-x 13 root root 4096 Apr 21 2018 ../ -rw-r--r-- 1 root root 0 Feb 15 2015 not_mounted
It seems to me that is _exactly_ the behaviour what you asked for when you specified 'nofail' = do not report if device does not exist.
Nope! The intended behaviour (to me) is not produce a crash when mounting during boot in an automated way. I'd like it to issue a message (fail) when done manually by me. Maybe not an error, but a warning. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
Carlos E. R. wrote:
On 22/02/2020 12.42, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
No, they don't. Thunderbird is typically configured the same way as Windows, which doesn't have an MTA. The client software simply talks directly with the MTA at the ISP.
As I said, it may be a matter of opinion. Personally, I want my cron daemon, smartd, mdadmd, apcupsd (and others like them) to be able to send emails. -- Per Jessen, Zürich (13.3°C) http://www.hostsuisse.com/ - virtual servers, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/02/2020 14.18, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 12.42, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
No, they don't. Thunderbird is typically configured the same way as Windows, which doesn't have an MTA. The client software simply talks directly with the MTA at the ISP.
As I said, it may be a matter of opinion. Personally, I want my cron daemon, smartd, mdadmd, apcupsd (and others like them) to be able to send emails.
Of course, and me too. Desktop operating systems, say Windows Home, Android, (Macs?) do not need an MTA. Linux does, which makes me say that Linux is a server class operating system. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
Carlos E. R. wrote:
On 22/02/2020 14.18, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 12.42, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
> To be fair, the current usual? environment of a single desktop > different considerably from a typical? server. If there's a > desktop, then what the (a?) user typically wants is for the > desktop to come up and then perhaps ask questions about what do > you want to do about this missing filesystem or whatever other > resource is missing. On a server, I might want it to stop and > bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
No, they don't. Thunderbird is typically configured the same way as Windows, which doesn't have an MTA. The client software simply talks directly with the MTA at the ISP.
As I said, it may be a matter of opinion. Personally, I want my cron daemon, smartd, mdadmd, apcupsd (and others like them) to be able to send emails.
Of course, and me too.
Desktop operating systems, say Windows Home, Android, (Macs?) do not need an MTA. Linux does, which makes me say that Linux is a server class operating system.
I was only commenting on your assertion "If I use postfix, that's clearly server software". Even my raspi runs postfix. MacOS is based on Unix, and most probably also has an MTA running. Maybe even postfix. I'll check, my wife has a Macbook. -- Per Jessen, Zürich (13.9°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/02/2020 15.13, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 14.18, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 12.42, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 08.24, Per Jessen wrote: > Dave Howorth wrote: > >> To be fair, the current usual? environment of a single desktop >> different considerably from a typical? server. If there's a >> desktop, then what the (a?) user typically wants is for the >> desktop to come up and then perhaps ask questions about what do >> you want to do about this missing filesystem or whatever other >> resource is missing. On a server, I might want it to stop and >> bleat "something's wrong", yes. > > Yes, that is a good point - I do tend to think mostly in terms of > servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
No, they don't. Thunderbird is typically configured the same way as Windows, which doesn't have an MTA. The client software simply talks directly with the MTA at the ISP.
As I said, it may be a matter of opinion. Personally, I want my cron daemon, smartd, mdadmd, apcupsd (and others like them) to be able to send emails.
Of course, and me too.
Desktop operating systems, say Windows Home, Android, (Macs?) do not need an MTA. Linux does, which makes me say that Linux is a server class operating system.
I was only commenting on your assertion "If I use postfix, that's clearly server software". Even my raspi runs postfix. MacOS is based on Unix, and most probably also has an MTA running. Maybe even postfix. I'll check, my wife has a Macbook.
Yes, postfix is server software :-) -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
On Sat, 22 Feb 2020 14:18:54 +0100 Per Jessen <per@computer.org> wrote:
Carlos E. R. wrote:
On 22/02/2020 12.42, Per Jessen wrote:
Carlos E. R. wrote:
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
No, they don't. Thunderbird is typically configured the same way as Windows, which doesn't have an MTA. The client software simply talks directly with the MTA at the ISP.
As I said, it may be a matter of opinion. Personally, I want my cron daemon, smartd, mdadmd, apcupsd (and others like them) to be able to send emails.
FWIW Raspis ship without an MTA by default. Even in the full desktop distribution. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Dave Howorth wrote:
On Sat, 22 Feb 2020 14:18:54 +0100 Per Jessen <per@computer.org> wrote:
Carlos E. R. wrote:
On 22/02/2020 12.42, Per Jessen wrote:
Carlos E. R. wrote:
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
No, they don't. Thunderbird is typically configured the same way as Windows, which doesn't have an MTA. The client software simply talks directly with the MTA at the ISP.
As I said, it may be a matter of opinion. Personally, I want my cron daemon, smartd, mdadmd, apcupsd (and others like them) to be able to send emails.
FWIW Raspis ship without an MTA by default. Even in the full desktop distribution.
I guess that is the Raspbian distro? I've never tried it, I run openSUSE on my Raspi. -- Per Jessen, Zürich (10.6°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 22 Feb 2020 18:58:32 +0100 Per Jessen <per@computer.org> wrote:
Dave Howorth wrote:
On Sat, 22 Feb 2020 14:18:54 +0100 Per Jessen <per@computer.org> wrote:
Carlos E. R. wrote:
On 22/02/2020 12.42, Per Jessen wrote:
Carlos E. R. wrote:
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Maybe a matter of opinion. Postfix is an MTA, every client machine needs one, desktops included. That is why we install it by default.
No, they don't. Thunderbird is typically configured the same way as Windows, which doesn't have an MTA. The client software simply talks directly with the MTA at the ISP.
As I said, it may be a matter of opinion. Personally, I want my cron daemon, smartd, mdadmd, apcupsd (and others like them) to be able to send emails.
FWIW Raspis ship without an MTA by default. Even in the full desktop distribution.
I guess that is the Raspbian distro? I've never tried it, I run openSUSE on my Raspi.
Yes, that's right. I use it because that's what all the h/w add-ons expect and give instructions for. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 22/02/2020 à 14:06, Carlos E. R. a écrit :
Nope! The intended behaviour (to me) is not produce a crash when mounting during boot in an automated way. I'd like it to issue a message (fail) when done manually by me. Maybe not an error, but a warning.
then the better option may be "noauto" jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/02/2020 14.51, jdd@dodin.org wrote:
Le 22/02/2020 à 14:06, Carlos E. R. a écrit :
Nope! The intended behaviour (to me) is not produce a crash when mounting during boot in an automated way. I'd like it to issue a message (fail) when done manually by me. Maybe not an error, but a warning.
then the better option may be "noauto"
It doesn't mount, but it tests for presence and aborts the boot if missing. Maybe if you also write "0 0" at the end of the line. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
Le 22/02/2020 à 17:42, Carlos E. R. a écrit :
On 22/02/2020 14.51, jdd@dodin.org wrote:
then the better option may be "noauto"
It doesn't mount, but it tests for presence and aborts the boot if missing. Maybe if you also write "0 0" at the end of the line.
that's odd... jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/02/2020 21.05, jdd@dodin.org wrote:
Le 22/02/2020 à 17:42, Carlos E. R. a écrit :
On 22/02/2020 14.51, jdd@dodin.org wrote:
then the better option may be "noauto"
It doesn't mount, but it tests for presence and aborts the boot if missing. Maybe if you also write "0 0" at the end of the line.
that's odd...
No, it has to test for need of fsck. man fstab The sixth field (fs_passno). This field is used by fsck(8) to determine the order in which filesystem checks are done at boot time. The root filesystem should be specified with a fs_passno of 1. Other filesystems should have a fs_passno of 2. Filesystems within a drive will be checked sequentially, but filesystems on different drives will be checked at the same time to utilize parallelism available in the hardware. Defaults to zero (don't fsck) if not present. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
jdd composed on 2020-02-22 21:05 (UTC+0100):
Carlos E. R. composed:
jdd wrote:
then the better option may be "noauto"
It doesn't mount, but it tests for presence and aborts the boot if missing. Maybe if you also write "0 0" at the end of the line.
that's odd...
I have many like this one, since well over a decade ago: LABEL=s131 /s131 ext3 noatime,noauto 0 0 -- Evolution as taught in public schools is religion, not science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 22 Feb 2020 11:35:34 +0100 "Carlos E. R." <robin.listas@telefonica.net> wrote:
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Whilst it's true that linux is designed to be used in servers, it's also true that linux is designed to be used as a desktop OS. But the whole discussion misses the point, which is that the desired behaviour of a machine that has a display and a user when encountering a fault is likely to be different to the desired behaviour of a server with no user in front of it. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
* Dave Howorth <dave@howorth.org.uk> [02-22-20 11:14]:
On Sat, 22 Feb 2020 11:35:34 +0100 "Carlos E. R." <robin.listas@telefonica.net> wrote:
On 22/02/2020 08.24, Per Jessen wrote:
Dave Howorth wrote:
To be fair, the current usual? environment of a single desktop different considerably from a typical? server. If there's a desktop, then what the (a?) user typically wants is for the desktop to come up and then perhaps ask questions about what do you want to do about this missing filesystem or whatever other resource is missing. On a server, I might want it to stop and bleat "something's wrong", yes.
Yes, that is a good point - I do tend to think mostly in terms of servers.
Many linux machines are also used as servers. Linux is really a server operating system, multiuser and multitasking. If I use postfix, that's clearly server software. A plain user simply doesn't make use of all the features.
Whilst it's true that linux is designed to be used in servers, it's also true that linux is designed to be used as a desktop OS.
But the whole discussion misses the point, which is that the desired behaviour of a machine that has a display and a user when encountering a fault is likely to be different to the desired behaviour of a server with no user in front of it.
perhaps but linux is not *just* a desktop but may be *just* a server, and you are free to adjust it to your desires. but will probably never achieve standing as an all-around system for all people even though it is much closer than anything else existing. -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Dave Howorth wrote:
But the whole discussion misses the point, which is that the desired behaviour of a machine that has a display and a user when encountering a fault is likely to be different to the desired behaviour of a server with no user in front of it.
I think you have a good point, never mind that I don't know what to do about it. -- Per Jessen, Zürich (9.9°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 22 Feb 2020 19:00:02 +0100 Per Jessen <per@computer.org> wrote:
Dave Howorth wrote:
But the whole discussion misses the point, which is that the desired behaviour of a machine that has a display and a user when encountering a fault is likely to be different to the desired behaviour of a server with no user in front of it.
I think you have a good point, never mind that I don't know what to do about it.
Well I'd guess that a better behaviour on a desktop m/c would be to come up to the desktop and then present a widget of some kind explaining what was wrong and offering possibilities for dealing with it (e.g. ignore it, fix it somehow, shut down, WHY?). As a second best, instead of failing the boot it could pause it and offer similar options whilst describing exactly what was wrong. Then retry the boot from that point after the USB stick was plugged in or it was marked as nofail or whatever. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/02/2020 20.51, Dave Howorth wrote:
On Sat, 22 Feb 2020 19:00:02 +0100 Per Jessen <> wrote:
Dave Howorth wrote:
But the whole discussion misses the point, which is that the desired behaviour of a machine that has a display and a user when encountering a fault is likely to be different to the desired behaviour of a server with no user in front of it.
I think you have a good point, never mind that I don't know what to do about it.
Well I'd guess that a better behaviour on a desktop m/c would be to come up to the desktop and then present a widget of some kind explaining what was wrong and offering possibilities for dealing with it (e.g. ignore it, fix it somehow, shut down, WHY?).
Can't do that. The system doesn't know if the desktop will fail because of the failed mount. And no, "/" and "/home" do not ensure a booting on every system.
As a second best, instead of failing the boot it could pause it and offer similar options whilst describing exactly what was wrong. Then retry the boot from that point after the USB stick was plugged in or it was marked as nofail or whatever.
Actually, that is what it does :-) It goes to emergency mode and gives you a cryptic message; you correct the issue, and issue ctrl-D to continue the boot process. -- Cheers / Saludos, Carlos E. R. (from 15.1 x86_64 at Telcontar)
Please note, that the problem at hand is already solved, but the underlying issue gets in my way from time to time, so I wanted to tell (or rant 😉) about it.
Two days ago, I wasn't able to boot my system (LEAP 15.1): X-Server didn't start and I could only log into maintenance mode with no network either. First, I suspected a botched update, but the kernel-firmware was updated a couple of days before and I had successfully booted in the mean time. So, I was dumbfounded, but then I remembered to look into /etc/fstab and found an entry of a USB drive that I had plugged in and removed. I deleted said entry and LEAP booted just fine.
My point is: maybe the maintainers could help to avoid this situation: Please, don't break the boot process if some malfunctioning but unimportant entry in fstab doesn't work! Or at least, make it easier for the user to find out what is going wrong. I had similar issues multiple times over the past years and sometimes it took me quite a time to figure them out.
Regards. I can confirm the issue if you use e.g. backintime and adminmode make the
In data venerdì 21 febbraio 2020 10:20:36 CET, zb4ng ha scritto: program take up the backuproutine when the disc is plugged. If the user shuts down during this, or if he forgets after the backup to umount, this will lead to such an entry (provided the usb disc is large and uses ext4). The machine will be unbootable and go into maintenance mode. _________________________________________________________________ ________________________________________________________ Ihre E-Mail-Postf�cher sicher & zentral an einem Ort. Jetzt wechseln und alte E-Mail-Adresse mitnehmen! https://www.eclipso.de -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 21.02.20 um 10:20 schrieb zb4ng:
Please note, that the problem at hand is already solved, but the underlying issue gets in my way from time to time, so I wanted to tell (or rant 😉) about it.
Two days ago, I wasn't able to boot my system (LEAP 15.1): X-Server didn't start and I could only log into maintenance mode with no network either. First, I suspected a botched update, but the kernel-firmware was updated a couple of days before and I had successfully booted in the mean time. So, I was dumbfounded, but then I remembered to look into /etc/fstab and found an entry of a USB drive that I had plugged in and removed. I deleted said entry and LEAP booted just fine.
My point is: maybe the maintainers could help to avoid this situation: Please, don't break the boot process if some malfunctioning but unimportant entry in fstab doesn't work! Or at least, make it easier for the user to find out what is going wrong. I had similar issues multiple times over the past years and sometimes it took me quite a time to figure them out.
Regards.
Just wanted to add that Ubuntu has the same behaviour. 1. create new partition in OpenSuSE 2. add partition to fstab in Ubuntu 3. delete partition in OpenSuSE 4. Ubuntu fails to boot -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (12)
-
Anton Aylward -
Carlos E. R. -
Dave Howorth -
Felix Miata -
James Knott -
jddï¼ dodin.org -
Patrick Shanahan -
Per Jessen -
Peter Suetterlin -
PGNet Dev -
stakanov -
zb4ng