[opensuse] btrfs (with subvolumes) & raid-1 setup
Hello, I'm considering to move from Debian (Sid) to openSuSe where my Linux adventure has begun in 1999. :-) At the moment I use raid-1 setup of 2 X 1TB GPT-partitioned disks with the following layout: /dev/sd(a,b)1 bios_grub 2MB /dev/sd(a,b)2 swap 8GB /dev/sd(a,b)3 btrfs 923.51GB /dev/sd(a,b)2 partitions are in raid1 swap volume, while btrfs partitions are also in raid-1 and I use subvolumes for / & /home: UUID=some-uuid / btrfs defaults,noatime,compress=lzo,autodefrag,subvol=@ 0 0 UUID=some-uuid /home btrfs defaults,noatime,compress=lzo,autodefrag,subvol=@home 0 0 I'd like to replicate this setup with Suse as well, but, if I'm right, the installer in 13.2 can't do it? The other option, which I used when installing Debian, is to install Suse on single disk and then clone it and add to raid-1 array. Any hint? Sincerely, Gour -- Bewildered by the modes of material nature, the ignorant fully engage themselves in material activities and become attached. But the wise should not unsettle them, although these duties are inferior due to the performers' lack of knowledge. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/18/2014 08:47 AM, Gour wrote:
I'd like to replicate this setup with Suse as well, but, if I'm right, the installer in 13.2 can't do it?
The other option, which I used when installing Debian, is to install Suse on single disk and then clone it and add to raid-1 array.
Any hint?
I don't know about the clone/raid; if it worked on Debian it will probably work on Suse. I *have* tried and succeeded at installing suse on a single disk with just /boot and swap as primary partitions and BtrFS for the final partition and nothing else. The difference between what you describe and what the installer led me too is * the ROOT FS is implicit, not a subvolume * I had subvolumes for /home, /usr, /var, /tmp and /.snapshots I later screwed things up by noticing that a subvolume was - functionally - like a subdirectory so why, I though, bother with the /etc/fstab entries? POOF! And no problems! So why have subvolumes at all if I'm not mounting them? POOF! Problems. They **ARE** like subdirectories. Remove subvolume is like "rm -fr" !! So I went back to my old way of doing things with LVM. But the "cant do it" is not so. At best you will have to tell the installer not to format the existing partition (and possibly not to try and create any more). But the partition manager part of the installer is quite capable in 'manual' mode. HOWEVER... Installing suse is going to overwrite your system. Isn't that the whole point? -- 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 Tue, 18 Nov 2014 09:02:47 -0500
Anton Aylward
So why have subvolumes at all if I'm not mounting them?
Well, I'm mounting them.
Installing suse is going to overwrite your system. Isn't that the whole point?
No, the point is that I'd dismantle raid-1 (mirror) setup, install Suse on one disk, keep my /home and then add 2nd disk to the array and sync the mirror. Sincerely, Gour -- Therefore, without being attached to the fruits of activities, one should act as a matter of duty, for by working without attachment one attains the Supreme. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Tue, 18 Nov 2014 09:02:47 -0500
Anton Aylward
* the ROOT FS is implicit, not a subvolume * I had subvolumes for /home, /usr, /var, /tmp and /.snapshots
Do you see any pro/cons in having ROOT FS as subvolume in regard to ability to have snapshots? (My current system has root as subvolume.) I'm also not sure if one can trick installer to do it... Sincerely, Gour -- Abandoning all attachment to the results of his activities, ever satisfied and independent, he performs no fruitive action, although engaged in all kinds of undertakings. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/22/2014 08:15 AM, Gour wrote:
On Tue, 18 Nov 2014 09:02:47 -0500 Anton Aylward
wrote: * the ROOT FS is implicit, not a subvolume * I had subvolumes for /home, /usr, /var, /tmp and /.snapshots
Do you see any pro/cons in having ROOT FS as subvolume in regard to ability to have snapshots? (My current system has root as subvolume.)
My opinion is this If the ROOT FS is the root then there's no sense it it also being a subvolume ... A subvolume of what? I've installed BtrFs as the ROOT FS for opensuse a few times not in a few variations including it being the only FS, the whole system with lots of subvolumes. It never occurred to me to have it as its own subvolume. And I've never had any problems with snapshots, for examples with zypper. The closest I've had to a "problem" with snapshots is when they were being taken on an hourly basis and eating disk space! Turn that setting OFF! -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 22 Nov 2014 09:23:22 -0500
Anton Aylward
If the ROOT FS is the root then there's no sense it it also being a subvolume
Well, the whole setup is cleaner and allows one to e.g. mount one's root to some other place, but I was not aware if there is some gotcha in regard to snapshots.
A subvolume of what?
Of btrfs filesystem. Sincerely, Gour -- The intricacies of action are very hard to understand. Therefore one should know properly what action is, what forbidden action is, and what inaction is. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/22/2014 10:03 AM, Gour wrote:
On Sat, 22 Nov 2014 09:23:22 -0500 Anton Aylward
wrote: If the ROOT FS is the root then there's no sense it it also being a subvolume
Well, the whole setup is cleaner and allows one to e.g. mount one's root to some other place, but I was not aware if there is some gotcha in regard to snapshots.
Each BtrFS has its own If, for example, as I once tried, you have /home in a separate container as a quite separate BtrFS you can mount it on /home on the root FS. It doesn't matter if the root FS is a BtrFS or not, and if it is it doesn't matter if the /home directory on the root FS is a subvolume or not as the mount of the external will, as is the case with any FS combination, overlay it. As it happened, when I tried that I _initially_ had the /home on the root at a subvolume, then unmounted the second drive, removed the subvolume, created a new /hone directory *which of course had nothing in it) and remounted /home. All was fine. You can configure the snapshot for each FS independently. The "cleaner" makes no sense to me. As for mount .... For quite another reason I once had to boot in maintenance mode and had a RAMFS and mounted my BtrFS containing the system on the RAMFS/mount, and did the other stuff needed so I could do a chroot. The RAMFS's fstab had no information about the BtrFS; the mount was from a manually entered command line. What I'm saying is that as far as mounting the 'ROOTFS' goes it being a BtrFS and having subvolumes MAKES NO DIFFERENCE AT ALL So there is no reason to have the root of the FS as a subvolume in the fstab.
A subvolume of what?
Of btrfs filesystem.
But it *IS* the root of the BtrFS. What you are trying to do is declare that the fs is subvolme of itself. The reality of subvolume-ness lies not in what is in the /etc/fstab but in what the FS thinks it has. You have the 12.3 installed? Try btrfs subvolume list -aqt / -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Отправлено с iPhone
22 нояб. 2014 г., в 16:15, Gour
написал(а): On Tue, 18 Nov 2014 09:02:47 -0500 Anton Aylward
wrote: * the ROOT FS is implicit, not a subvolume * I had subvolumes for /home, /usr, /var, /tmp and /.snapshots
Do you see any pro/cons in having ROOT FS as subvolume in regard to ability to have snapshots? (My current system has root as subvolume.)
I think this the only setup that makes sense in the long run. I do not know if yast supports it though.
I'm also not sure if one can trick installer to do it...
Sincerely, Gour
-- Abandoning all attachment to the results of his activities, ever satisfied and independent, he performs no fruitive action, although engaged in all kinds of undertakings.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 22 Nov 2014 17:38:38 +0300 arvidjaar@gmail.com wrote:
I think this the only setup that makes sense in the long run. I do not know if yast supports it though.
It seems it does not, so one would have to rsync after the install is complete. Sincerely, Gour -- Thus the wise living entity's pure consciousness becomes covered by his eternal enemy in the form of lust, which is never satisfied and which burns like fire. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/22/2014 09:38 AM, arvidjaar@gmail.com wrote:
I think this the only setup that makes sense in the long run. I do not know if yast supports it though.
Please explain why you feel it is the only setup that makes sense. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
В Sat, 22 Nov 2014 13:34:16 -0500
Anton Aylward
On 11/22/2014 09:38 AM, arvidjaar@gmail.com wrote:
I think this the only setup that makes sense in the long run. I do not know if yast supports it though.
Please explain why you feel it is the only setup that makes sense.
You install new updates and create snapshot before it. Your primary filesystem is in btrfs root and snapshot goes in some subvolume (/.snapshot/backup or anything). You reboot and find that your system does not boot or is not usable. What are your options to go back to working backup? 1. Copy files from backup snapshot back into root. It is time consuming, defeats space saving by inflating root and may not be that trivial if directory structure changed. 2. Mount backup snapshot as root. So we are exactly in the situation we discuss. We end up with root on subvolume without explicit support from system management tools. 3. Set default volume to backup snapshot Probably just as unsupported as 2; also it raises hairy questions like "what / now refers to, like in /.snapshot". If your root is on subvolume from the very beginning and all tools are known to support it, fallback is trivial. You are on /@/root-current, you create /@/root-backup; if boot from /@/root-current fails, just point to /@/root-backup and continue to work. Second consideration is that "backup before install" is backward. It is not safe to update running system, even if update looks very innocent. The only sure way is to clone current system, update clone (while running your current system) and reboot into clone. And current model simply does not provide for it. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 01:31 AM, Andrei Borzenkov wrote:
В Sat, 22 Nov 2014 13:34:16 -0500 Anton Aylward
пишет: On 11/22/2014 09:38 AM, arvidjaar@gmail.com wrote:
I think this the only setup that makes sense in the long run. I do not know if yast supports it though.
Please explain why you feel it is the only setup that makes sense.
You install new updates and create snapshot before it. Your primary filesystem is in btrfs root and snapshot goes in some subvolume (/.snapshot/backup or anything). You reboot and find that your system does not boot or is not usable. What are your options to go back to working backup?
That's an interesting strawman argument but I don't see it as being realistic or common. And it has nothing whatsoever to do with root as its own subvolume.
1. Copy files from backup snapshot back into root.
It is time consuming, defeats space saving by inflating root and may not be that trivial if directory structure changed.
You seem confused as to the purpose of a snapshot and and confused as to why you should want to reboot after a zypper update. The only time you need to update following a 'update' is of you have replaced or rebuilt the kernel. That is a special case.
2. Mount backup snapshot as root.
So we are exactly in the situation we discuss. We end up with root on subvolume without explicit support from system management tools.
That is not so. There is a mistake in your thinking. Sometimes I get a DUH reaction reading this list and then I realise that not everyone has had the opportunity to be exposed to "Big Iron administrative practices. Snapshoting of the 'before' state when applying changes date back to Big Iron. Snapshots are about being able to roll back updates. They are not backups. We have snapshotting on BtrFS but not on, for example, ReiserFS, because of the COW mechanism if you walk, for example, the /etc tree and the /.snapshot/<number>/snapshot/etc/ tree side by side you will see that for each file they appear to be hardlinked. Actually the snapshot version ow COWed. If it really were a backup then it would be a copy, and making a full copy of the file system for every snapshot will rapidly consume your disk. The state info is bad enough. The purpose of a snapshot is so you can roll back the change(s) resulting from a zypper update. Or possibly, of you have regular by the clock snapshotting enabled, from changes you make by had to the configuration.. let me repeat: A SNAPSHOT IS NOT A BACKUP If you have badly screwed what is in /etc and rendered your system unbootable because of that that mounting a snapshot is not a fix.
3. Set default volume to backup snapshot
Probably just as unsupported as 2; also it raises hairy questions like "what / now refers to, like in /.snapshot".
If your root is on subvolume from the very beginning and all tools are known to support it, fallback is trivial. You are on /@/root-current, you create /@/root-backup; if boot from /@/root-current fails, just point to /@/root-backup and continue to work.
That is completely unjustified and not logical. Firstly in recovery mode there is no need for the root of the file system you are 'mending' to be a subvolume. As I've said before, the 'mending' from recovery of an unbootable systems is a well established procedure and works with any file system. It works, I know from my own experience, with BtrFS without the need for any subvolume trickery. Having booted from the recovery CD/DVD of your choice you can mount the hard disk RootFS at /mnt. You can do this with ANY file system, ext[234], ReiserFS, XFS or BtrFS. All the tools available on your recovery system are still available. If it is a screwed system config that is the reason your system is unbootable then the snapshots look just like a subdirectory and you'll have to fire which one has the prior actual copy. My experience with 'unbootable' has more to do with what's in /boot. Fixing that is quite another matter. It also is independent of the file system being used. All in all you have not presented an argument why a BtrFS has to be a subvolume of itself. The arguments you give are not about 'necessity; what you are trying to achieve with these 'mounts' can be done without the need for subvolumes. Snapshots with BtrFS are an emergent property of the COW mechanism. They are not copies in the conventional sense. Their purpose is for rolling back changes, not for recovery from catastrophic disaster. If you are using the COW mechanism as a substitute for proper backups -- as seems to be the case with the scenarios you describe -- then you are mistaken and going to complicate your life if you do have a disaster that necessitates recovery. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 10:23:35 -0500
Anton Aylward
That's an interesting strawman argument but I don't see it as being realistic or common.
I suggest you to visit #btrfs (Freenode) and/or some btrfs mailing lists to expand your vision a bit. ;) Sincerely, Gour -- Bewildered by the modes of material nature, the ignorant fully engage themselves in material activities and become attached. But the wise should not unsettle them, although these duties are inferior due to the performers' lack of knowledge. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 10:53 AM, Gour wrote:
On Sun, 23 Nov 2014 10:23:35 -0500 Anton Aylward
wrote: That's an interesting strawman argument but I don't see it as being realistic or common.
I suggest you to visit #btrfs (Freenode) and/or some btrfs mailing lists to expand your vision a bit. ;)
You mistake my meaning and intent. Which was that this has anything to do with the necessity of a) boot failures being specific to BtrFS systems b) the options available for "boot debugging" that are applicable to other file systems are somehow inapplicable to BtrFS There's still been noting state that necessitates having the root of a BtrFS being a subvolume of itself. Those arguments too are all strawmen, Please do not construe this to mean that there are not ways to render a system unbootable, for various meanings of 'unbootable', nor that those ways can be carried out on a system with a BrtrFS root file system. BTDT. Nor does it mean that earlier versions of the BtrFS were unstable enough to cause the systems to become unbootable for reasons that were specific to the root BtrFS. Neither does that mean that even mature file systems do not have fringe conditions, some discovered, some yet to be discovered, that can make them unstable. The reality is that BtrFS is the youngest of the 'popular' file systems because of its association with SSD and has been, like systemd, like KDE4, pushed out aggressively even while immature, and has garnered q lot of criticism when like Kde4 and systemd, its early instances met with problems and failures. Such is the nature of the FSS development model. I've been using BtrFS without BtrFs specific boot problems since late opensuse 12.1. Yes, I've had kernel updates that have gone wrong, but the 'have gone wrong' was confined to /boot which is a ext2FS. The technique I described of using the recovery disk to mount a BtrFS rootFS on the RAMFS /mnt, then mount the hard disk partition /boot on /mnt/boot, then do the /sys, /dev, /proc trickery and chroot to /mnt and then mkinitrd all works. It worked before I started using BtrFS and it works now on a BtrFS without a subvolume of itself. Hypothesising a elaborate configuration specifically to 'show' that a subvolume-of-itself is needed is an unrealistic strawman. I'm not sure how I would classify failing to recognise that snapshots are based around COW rather than real copies. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 11:25:53 -0500
Anton Aylward
There's still been noting state that necessitates having the root of a BtrFS being a subvolume of itself. Those arguments too are all strawmen,
You're mixing apples & oranges... First of all I *never* wrote that 'having the root of btrfs as subvolume' is *necessity* - it's just convenience. Secondly, my system boots and the problem is how to configure Grub2 to do it without my intervention - if you managed to read my reply to Andrei, you can see what's the problem. Iow, the problem is some bug in Suse's grub-mkconfig, since on Debian I get something like: [...] linux /@/boot/vmlinuz-3.16-2-amd64 root=xyz ro rootflags=subvol=@ quiet initrd /@/boot/initrd.img-3.16-2-amd64 and this should be fixed!! (I already used my free-time slots today, so if don't get some reply from Andrei, will take a closer look myself.) Sincerely, Gour -- A person is said to be elevated in yoga when, having renounced all material desires, he neither acts for sense gratification nor engages in fruitive activities. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 12:05 PM, Gour wrote:
Iow, the problem is some bug in Suse's grub-mkconfig, since on Debian I get something like:
No, its not a problem. The Debian way, I gather, is to have that /@/ The openSuse way is not to. As far as I can see the problem is that rather than accept that each distribution has its own way of doing things you are, as you say
I want is to have *identical* setup to which I'm accustomed
That is you want to make the openSuse look like a Debian system. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 12:40:10 -0500
Anton Aylward
The Debian way, I gather, is to have that /@/ The openSuse way is not to.
1. Having btrfs root under subvolume is not anything Debian-specific and it's actually not supported by Debian installer. You can find tutorials all over the net explaining how to do it for *many* Linux distros and it's even mentioned in Suse's forums.
As far as I can see the problem is that rather than accept that each distribution has its own way of doing things you are, as you say
2. Every distribution has its own installer which covers *many* use-cases, but certainly *not all* the use cases. For instance, I have raid-1 setup on Debian which is not supported by its installer nor it's supported by Yast/Suse. By following your logic, user should not put two hard disks in raid-1 array 'cause it's not supported by the distro - here I think not about using mdadm, but btrfs' own raid capabilities - which is, of course, non sensical.
That is you want to make the openSuse look like a Debian system.
3. As I already wrote, disk layout is nothing which is distro-specific and, at the end, Debian, Suse, Ubuntu, Fedora...that's all *same* OS called Linux which is known to provide 'choices' to the end-users. By your logic, all distro-users should have their systems configured the *same* way. :-) Sincerely, Gour -- Never was there a time when I did not exist, nor you, nor all these kings; nor in the future shall any of us cease to be. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, Nov 23, 2014 at 6:23 PM, Anton Aylward
That's an interesting strawman argument
I do not argue with anything. You asked me to explain and I explained.
Firstly in recovery mode there is no need for the root of the file system you are 'mending' to be a subvolume. As I've said before, the 'mending' from recovery of an unbootable systems is a well established procedure and works with any file system. It works, I know from my own experience, with BtrFS without the need for any subvolume trickery.
Having booted from the recovery CD/DVD of your choice you can mount the hard disk RootFS at /mnt. You can do this with ANY file system, ext[234], ReiserFS, XFS or BtrFS.
All the tools available on your recovery system are still available.
Production systems have limited maintenance window when changes can be applied. If anything goes wrong they do not have luxury of spending time to recover - you need system up and running as soon as possible. btrfs opens up possibility of having cheap and space efficient fallback to known good state so you can easily revert changes. Whether you are going to use this possibility or not - is up to you. I think this possibility is very useful and that current implementation in (open)SUSE makes it less convenient to use than it could be. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 04:01 PM, Andrei Borzenkov wrote:
Production systems have limited maintenance window when changes can be applied. If anything goes wrong they do not have luxury of spending time to recover - you need system up and running as soon as possible. btrfs opens up possibility of having cheap and space efficient fallback to known good state so you can easily revert changes. Whether you are going to use this possibility or not - is up to you. I think this possibility is very useful and that current implementation in (open)SUSE makes it less convenient to use than it could be.
Where time is critical there are other mechanism such as 'hot cut=over'. What's missing from this part of the thread is the specifics. Not all catastrophic instances are the same. Change management, that is apply changes and updates, is not as simple for critical production systems such as banks, telecos, brokerages and the like. I know this first-hand having done my time on the change control committee for a national bank. Companies that have the capability test out changes on non-critical systems before applying them to critical systems. That goes just as much for "mainframes" as for armies of Pcs. Yes, the COW mechanism we now have makes reversing changes much easier, and opens up alternatives for organizations that don't have the back-room of test-bed systems and staff to exercise the hell out of updates. But it does mean that the business and operational processes involved in applying updates to critical systems have to be worked out. While I can on my non-critical, or at least non time critical, system apply 'zypper up' daily, and can do that 99.9% of the time without having to concern myself about reversing a update, I think that a business which applies all available updates willy-nilly is foolhardy. Relying on snapshotting to be able to reverse changes unconditionally is optimistic. A proper change control procedure will examine each individual update to see if it or one of the changes it implies will affect any of the production processes and if so how. Realistically we fond that many were obvious and straight forward and easy to verify.It helps when the change references a CVE. One of the advantages of Linux over a big IBM 3xxx or whatever class they are calling mainframes now is that changes can be tested on a comparatively cheap machine :-) The BtrFS snapshotting revolves around COW. COW is not a true copy, what's in a snapshot is not a real copy of the 'prior state'. It is not a backup. Don't try treating it as one. You can't take it away like you can a backup to tape or DVD or USB. And its not like a LVM2 snapshot either! There are ways to crash a file system where the 'main' file is corrupted but the COW copy has not been able to be ... Well, copied. bear that in mind too. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 22 Nov 2014 17:38:38 +0300 arvidjaar@gmail.com wrote:
I think this the only setup that makes sense in the long run. I do not know if yast supports it though.
I did rename the default subvolumes and created snapshot of rootfs in order to move it from top-level into subvolume (e.g. '@'), but have problem how to teach Yast/Grub to generate proper paths, iow. how to use /@/boot/... instead of /boot/... since manually editing during the boot time works? Any hint? Sincerely, Gour -- Whatever action a great man performs, common men follow. And whatever standards he sets by exemplary acts, all the world pursues. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/22/2014 03:29 PM, Gour wrote:
I did rename the default subvolumes and created snapshot of rootfs in order to move it from top-level into subvolume (e.g. '@'),
Yes, but WHY? I've got snapshots of the rootFS, and changes to the root FS, all without doing that.
but have problem how to teach Yast/Grub to generate proper paths, iow. how to use
/@/boot/... instead of /boot/... since manually editing during the boot time works?
Perhaps you should take the fact that yast can't/wont do that as a hint that it shouldn't/needn't be done? -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
В Sat, 22 Nov 2014 21:29:56 +0100
Gour
On Sat, 22 Nov 2014 17:38:38 +0300 arvidjaar@gmail.com wrote:
I think this the only setup that makes sense in the long run. I do not know if yast supports it though.
I did rename the default subvolumes and created snapshot of rootfs in order to move it from top-level into subvolume (e.g. '@'), but have problem how to teach Yast/Grub to generate proper paths, iow. how to use
/@/boot/... instead of /boot/... since manually editing during the boot time works?
Any hint?
Could you remind which openSUSE version you run? Could you be more specific where it fails? There are three places that need adjustment 1. reference to /boot/grub2 in first stage grub binary. If /boot/grub2 was on the same subvolume that was moved (and is now /@/boot/grub2) you need to rerun grub2-install so it picks up new path. 2. Paths to kernels in grub.cfg. They should be updated after rerunning grub2-mkconfig. If not, it is a bug and I would be interested in more information. 3. kernel command line options, specifically subvol=.... Here I'm, not sure. There code to do it is there, but I do not know whether it works. So what exactly does not work in your case? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 01:39 AM, Andrei Borzenkov wrote:
So what exactly does not work in your case?
I think he is just strawmaning. Certainly the scenario he describes is vague and your specific questions make sense. But normal good practice is to put /boot on its own partition, an ext2 or ext3 seems to be the consensus, and that takes it out of the realm of failure modes of BtrFS and snapshotting. Assuming you don't have a hard disk failure, which, sadly, is an all to common event these days, as we've discussed the reasons for elsewhere on this forum, then the likelihood of a boot failure is going to have at it cause some change to what is in /boot. I say that because there are failure modes that will get the systems up but only to a command prompt, single user mode, and that's quite another matter. I realise some people call 'not going into the gui' a boot failure. I don't. A boot failure will result from a) manually running 'mkinitrd' or similar and getting it wrong b) installing a new kernel, manually or automatically, and it going wrong c) doing an installation and it going wrong. None of those are specific to BtrFS. The fix for case (c) is pretty much "do it over". The fix for (a) and (b) ... Google for "Boot Debugging". As I say, it is a pretty well established procedure and has nothing to do with BtrFS as opposed to any other file system, so the issue of having the root of the file system as s subvolume of itself is quite irrelevant. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 09:39:56 +0300
Andrei Borzenkov
Could you remind which openSUSE version you run?
13.2
Could you be more specific where it fails? There are three places that need adjustment
OK. I managed to rename original subvolumes and put root from the top-level into subvolume. Here is my /etc/fstab: UUID=xyz / btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@ 0 0 UUID=xyz /home btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@home 0 0 UUID=xyz /opt btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@opt 0 0 UUID=xyz /srv btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@srv 0 0 UUID=xyz /tmp btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@tmp 0 0 UUID=xyz /usr/local btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@usr_local 0 0 UUID=xyz /var/crash btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@var_crash 0 0 UUID=xyz /var/log btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@var_log 0 0 UUID=xyz /var/opt btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@var_opt 0 0 UUID=xyz /var/spool btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@var_spool 0 0 UUID=xyz /var/tmp btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@var_tmp 0 0 UUID=xyz swap swap defaults 0 0 UUID=xyz /.snapshots btrfs noatime,noatime,autodefrag,compress=lzo,subvol=@snapshots 0 0 However, I have problem with Grub...
1. reference to /boot/grub2 in first stage grub binary. If /boot/grub2 was on the same subvolume that was moved (and is now /@/boot/grub2) you need to rerun grub2-install so it picks up new path.
Here is the result: # grub2-install /dev/sda Installing for i386-pc platform. Installation finished. No error reported. I notice that /boot/grub2(x86_64-efi directory is empty and wonder if it's normal considering that my install is on x86_64 with BIOS/GPT disk?
2. Paths to kernels in grub.cfg. They should be updated after rerunning grub2-mkconfig. If not, it is a bug and I would be interested in more information.
This is buggy and here are the details with my inline comments: [...] set root='hd0,gpt3' # path is set correctly [...] menuentry 'openSUSE' [...] linux /boot/vmlinuz-3.16.6-2-desktop root=UUID=xyz ${extra_cmdline} resume=/dev/disk/by-uuid/xyz splash=silent quiet showopts # the path is missing /@/boot/ as well as rootflags=subvol=@ echo 'Loading initial ramdisk ...' initrd /boot/initrd-3.16.6-2-desktop # same here - wrong path which should be /@/boot/... [...] menuentry 'openSUSE, with Linux 3.16.6-2-desktop (recovery mode)' [...] linux /boot/vmlinuz-3.16.6-2-desktop root=UUID=xyz ${extra_cmdline} showopts apm=off noresume edd=off powersaved=off nohz=off highres=off processor.max_cstate=1 nomodeset x11failsafe rootflags=subvol=@ # it's strange that here there is rootflags?? but the path is still # wrong echo 'Loading initial ramdisk ...' initrd /boot/initrd-3.16.6-2-desktop # again, wrong path insmod gfxmenu loadfont ($root)/boot/grub2/themes/openSUSE/ascii.pf2 loadfont ($root)/boot/grub2/themes/openSUSE/DejaVuSans10.pf2 loadfont ($root)/boot/grub2/themes/openSUSE/DejaVuSans12.pf2 loadfont ($root)/boot/grub2/themes/openSUSE/DejaVuSans-Bold14.pf2 insmod png set theme=($root)/boot/grub2/themes/openSUSE/theme.txt # the above themes-paths are also wrong!!
3. kernel command line options, specifically subvol=.... Here I'm, not sure. There code to do it is there, but I do not know whether it works.
I hope, from the above it's clear what's wrong? Sincerely, Gour -- In this endeavor there is no loss or diminution, and a little advancement on this path can protect one from the most dangerous type of fear. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 16:49:43 +0100
Gour
I hope, from the above it's clear what's wrong?
Forgot to wrote, that in order to boot I have to repeat (ad infinitum) the following sequence after Grub stops in th rescue mode:
set prefix=(hd0,gpt3)/@/boot/grub2 set root=(hd0,gpt3) insmod normal normal
which gives me opportunity to fix/edit 'linux' & 'initrd' lines after which openSuse-13.2 boots normally. Sincerely, Gour -- Before giving up this present body, if one is able to tolerate the urges of the material senses and check the force of desire and anger, he is well situated and is happy in this world. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, Nov 23, 2014 at 7:01 PM, Gour
On Sun, 23 Nov 2014 16:49:43 +0100 Gour
wrote: I hope, from the above it's clear what's wrong?
Forgot to wrote, that in order to boot I have to repeat (ad infinitum) the following sequence after Grub stops in th rescue mode:
set prefix=(hd0,gpt3)/@/boot/grub2 set root=(hd0,gpt3) insmod normal normal
which gives me opportunity to fix/edit 'linux' & 'initrd' lines after which openSuse-13.2 boots normally.
If it fails to compute correct path it is just expected that it fails everywhere. Could you attach /proc/self/mountinfo as well as output of btrfs subvolume list -a / btrfs subvolume get-default / grub2-mkrelpath / grub2-mkrelpath /boot/grub2 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 23:48:45 +0300
Andrei Borzenkov
Could you attach /proc/self/mountinfo as well as output of
# cat /proc/self/mountinfo 15 35 0:15 / /sys rw,nosuid,nodev,noexec,relatime shared:6 - sysfs sysfs rw 16 35 0:3 / /proc rw,nosuid,nodev,noexec,relatime shared:5 - proc proc rw 17 35 0:5 / /dev rw,nosuid shared:2 - devtmpfs devtmpfs rw,size=501528k,nr_inodes=125382,mode=755 18 15 0:10 / /sys/kernel/security rw,nosuid,nodev,noexec,relatime shared:7 - securityfs securityfs rw 19 17 0:16 / /dev/shm rw,nosuid,nodev shared:3 - tmpfs tmpfs rw 20 17 0:12 / /dev/pts rw,nosuid,noexec,relatime shared:4 - devpts devpts rw,gid=5,mode=620,ptmxmode=000 21 35 0:17 / /run rw,nosuid,nodev shared:20 - tmpfs tmpfs rw,mode=755 22 15 0:18 / /sys/fs/cgroup rw,nosuid,nodev,noexec shared:8 - tmpfs tmpfs rw,mode=755 23 22 0:19 / /sys/fs/cgroup/systemd rw,nosuid,nodev,noexec,relatime shared:9 - cgroup cgroup rw,xattr,release_agent=/usr/lib/systemd/systemd-cgroups-agent,name=systemd 24 15 0:20 / /sys/fs/pstore rw,nosuid,nodev,noexec,relatime shared:19 - pstore pstore rw 25 22 0:21 / /sys/fs/cgroup/cpuset rw,nosuid,nodev,noexec,relatime shared:10 - cgroup cgroup rw,cpuset 26 22 0:22 / /sys/fs/cgroup/cpu,cpuacct rw,nosuid,nodev,noexec,relatime shared:11 - cgroup cgroup rw,cpu,cpuacct 27 22 0:23 / /sys/fs/cgroup/memory rw,nosuid,nodev,noexec,relatime shared:12 - cgroup cgroup rw,memory 28 22 0:24 / /sys/fs/cgroup/devices rw,nosuid,nodev,noexec,relatime shared:13 - cgroup cgroup rw,devices 29 22 0:25 / /sys/fs/cgroup/freezer rw,nosuid,nodev,noexec,relatime shared:14 - cgroup cgroup rw,freezer 30 22 0:26 / /sys/fs/cgroup/net_cls,net_prio rw,nosuid,nodev,noexec,relatime shared:15 - cgroup cgroup rw,net_cls,net_prio 31 22 0:27 / /sys/fs/cgroup/blkio rw,nosuid,nodev,noexec,relatime shared:16 - cgroup cgroup rw,blkio 32 22 0:28 / /sys/fs/cgroup/perf_event rw,nosuid,nodev,noexec,relatime shared:17 - cgroup cgroup rw,perf_event 33 22 0:29 / /sys/fs/cgroup/hugetlb rw,nosuid,nodev,noexec,relatime shared:18 - cgroup cgroup rw,hugetlb 35 0 0:31 /@ / rw,noatime shared:1 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 36 16 0:34 / /proc/sys/fs/binfmt_misc rw,relatime shared:21 - autofs systemd-1 rw,fd=26,pgrp=1,timeout=300,minproto=5,maxproto=5,direct 37 17 0:14 / /dev/mqueue rw,relatime shared:22 - mqueue mqueue rw 38 15 0:6 / /sys/kernel/debug rw,relatime shared:23 - debugfs debugfs rw 39 17 0:35 / /dev/hugepages rw,relatime shared:24 - hugetlbfs hugetlbfs rw 40 35 0:31 /@snapshots /.snapshots rw,noatime shared:25 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 42 35 0:31 /@var_tmp /var/tmp rw,noatime shared:26 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 44 35 0:31 /@tmp /tmp rw,noatime shared:27 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 46 35 0:31 /@usr_local /usr/local rw,noatime shared:28 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 48 35 0:31 /@srv /srv rw,noatime shared:29 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 50 35 0:31 /@var_log /var/log rw,noatime shared:30 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 52 35 0:31 /@var_opt /var/opt rw,noatime shared:31 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 54 35 0:31 /@var_spool /var/spool rw,noatime shared:32 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 58 35 0:31 /@var_crash /var/crash rw,noatime shared:33 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 56 35 0:31 /@home /home rw,noatime shared:34 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 60 35 0:31 /@opt /opt rw,noatime shared:35 - btrfs /dev/sda3 rw,compress=lzo,space_cache,autodefrag 89 21 0:48 / /run/user/1000/gvfs rw,nosuid,nodev,relatime shared:73 - fuse.gvfsd-fuse gvfsd-fuse rw,user_id=1000,group_id=100
btrfs subvolume list -a /
# btrfs subvolume list -a /
ID 271 gen 353 top level 5 path
btrfs subvolume get-default /
# btrfs subvolume get-default / ID 5 (FS_TREE)
grub2-mkrelpath / # grub2-mkrelpath / /@
grub2-mkrelpath /boot/grub2 # grub2-mkrelpath /boot/grub2 /@/boot/grub2
Of course, the above output was produced in booted Suse-13.2 after I performed the:
set prefix=(hd0,gpt3)/@/boot/grub2 set root=(hd0,gpt3) insmod normal normal
sequence. Sincerely, Gour -- While contemplating the objects of the senses, a person develops attachment for them, and from such attachment lust develops, and from lust anger arises. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Mon, Nov 24, 2014 at 11:55 AM, Gour
grub2-mkrelpath / # grub2-mkrelpath / /@
grub2-mkrelpath /boot/grub2 # grub2-mkrelpath /boot/grub2 /@/boot/grub2
Absolutely fine.
Of course, the above output was produced in booted Suse-13.2 after I performed the:
set prefix=(hd0,gpt3)/@/boot/grub2 set root=(hd0,gpt3) insmod normal normal
sequence.
You mean that if you *now* run grub2-mkconfig and/or grub2-install they will produce paths without /@? Could you please attach output of grub2-mkconfig and grub2-install --debug? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Mon, 24 Nov 2014 12:03:56 +0300
Andrei Borzenkov
You mean that if you *now* run grub2-mkconfig and/or grub2-install they will produce paths without /@?
Yes.
Could you please attach output of grub2-mkconfig
See mkconfig.txt in the attachment.
and grub2-install --debug?
See grub2install.txt.bz2 Let me say: "Thank you!" for your readiness to help instead of speaking philosophy. ;) Sincerely, Gour -- A self-realized man has no purpose to fulfill in the discharge of his prescribed duties, nor has he any reason not to perform such work. Nor has he any need to depend on any other living being.
On Mon, Nov 24, 2014 at 12:51 PM, Gour
On Mon, 24 Nov 2014 12:03:56 +0300 Andrei Borzenkov
wrote: You mean that if you *now* run grub2-mkconfig and/or grub2-install they will produce paths without /@?
Yes.
Weird. I'll need to reproduce it, but I'm on business trip this week without access to my system so I'll be able to do it on weekend. Sorry. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Mon, 24 Nov 2014 13:14:59 +0300
Andrei Borzenkov
Weird. I'll need to reproduce it, but I'm on business trip this week without access to my system so I'll be able to do it on weekend. Sorry.
The crucial thing to solve was to do: btrfs subvolume set-default id-of-@ / Then there was no more resuce mode in Grub and I found out that eve Yast/Boot_Loader generates correct linux/initrd lines - not using '@' in path, but it, somehow, works. :-) Otoh, Debian puts '@' in the paths, but does not require to set-default. Sincerely, Gour -- One who works in devotion, who is a pure soul, and who controls his mind and senses is dear to everyone, and everyone is dear to him. Though always working, such a man is never entangled. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Mon, Nov 24, 2014 at 2:54 PM, Gour
On Mon, 24 Nov 2014 13:14:59 +0300 Andrei Borzenkov
wrote: Weird. I'll need to reproduce it, but I'm on business trip this week without access to my system so I'll be able to do it on weekend. Sorry.
The crucial thing to solve was to do:
btrfs subvolume set-default id-of-@ /
It should not be necessary. Actually, grub was explicitly changed to not require set-default. Good that you have workaround, but I will look at it.
Then there was no more resuce mode in Grub and I found out that eve Yast/Boot_Loader generates correct linux/initrd lines - not using '@' in path, but it, somehow, works. :-)
I'm afraid it has something to do with suse-specific patches ... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Mon, 24 Nov 2014 15:28:35 +0300
Andrei Borzenkov
It should not be necessary. Actually, grub was explicitly changed to not require set-default. Good that you have workaround, but I will look at it.
Let me say that I did repeat the same receipt with another install on my netbook and can confirm that it works.
I'm afraid it has something to do with suse-specific patches ...
I'd also like to add that, imho, installer should get the option to provide ability to define fstab parameters for *every* subvolume which would make the whole procedure a snap. At the end, I like Suse' installer and after finishing backup of my desktop machine, I'm going towards full migration to Suse/KDE. ;) Sincerely, Gour -- An intelligent person does not take part in the sources of misery, which are due to contact with the material senses. O son of Kuntī, such pleasures have a beginning and an end, and so the wise man does not delight in them. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
В Mon, 24 Nov 2014 15:28:35 +0300
Andrei Borzenkov
On Mon, Nov 24, 2014 at 2:54 PM, Gour
wrote: On Mon, 24 Nov 2014 13:14:59 +0300 Andrei Borzenkov
wrote: Weird. I'll need to reproduce it, but I'm on business trip this week without access to my system so I'll be able to do it on weekend. Sorry.
The crucial thing to solve was to do:
btrfs subvolume set-default id-of-@ /
It should not be necessary. Actually, grub was explicitly changed to not require set-default. Good that you have workaround, but I will look at it.
Then there was no more resuce mode in Grub and I found out that eve Yast/Boot_Loader generates correct linux/initrd lines - not using '@' in path, but it, somehow, works. :-)
I'm afraid it has something to do with suse-specific patches ...
Yes, it is SUSE specific patch to facilitate snapshot-booting. If SUSE_BTRFS_SNAPSHOT_BOOTING is set to true in /etc/default/grub (I think, it is default during new installation), grub2-mkconfig will generate paths relative to SUBVOLUME, not full absolute paths as upstream does. So e.g. for /some/subvol/path/to/file it will produce just /path/to/file. The intended usage is - you have root in top level subvolume. For this case subvolume is empty and paths are correct - you have snapshots in some subvolume(s). To boot from snapshot, you set grub variable btrfs_subvol to the subvolume path. This is supposed to be done automatically by grub scripts. So paths in snapshot remain the same (/boot/vmlinuz) but now point to files in different subvolume. Where it fails, is the case of primary root on subvolume - it never generates those magic variables to point to to it. I remember speaking with mchang and he mentioned that case of having root on separate subvolume was never intended to be supported. Unsetting SUSE_BTRFS_SNAPSHOT_BOOTING (or setting it to false) will revert to upstream behavior. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 10:49 AM, Gour wrote:
Here is my /etc/fstab:
Just to make sure: You do not have a separate /boot listed there. I'd be interested in learning why you chose not to have a separate /boot partition. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 11:29:02 -0500
Anton Aylward
You do not have a separate /boot listed there.
Correct.
I'd be interested in learning why you chose not to have a separate /boot partition.
Because the Linux OS has advanced and there is no longer required to keep separate /boot. Moreover, I use Debian with such setup for more than one year without any problem: UUID=xyz / btrfs defaults,noatime,compress=lzo,autodefrag,subvol=@ 0 0 UUID=xyz /home btrfs defaults,noatime,compress=lzo,autodefrag,subvol=@home 0 0 As you can see, there are not so many subvolumes as created by default by Yast, but it works. Sincerely, Gour -- A person who has given up all desires for sense gratification, who lives free from desires, who has given up all sense of proprietorship and is devoid of false ego — he alone can attain real peace. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 11:46 AM, Gour wrote:
On Sun, 23 Nov 2014 11:29:02 -0500 Anton Aylward
wrote: You do not have a separate /boot listed there.
Correct.
I'd be interested in learning why you chose not to have a separate /boot partition.
Because the Linux OS has advanced and there is no longer required to keep separate /boot.
Moreover, I use Debian with such setup for more than one year without any problem:
UUID=xyz / btrfs defaults,noatime,compress=lzo,autodefrag,subvol=@ 0 0 UUID=xyz /home btrfs defaults,noatime,compress=lzo,autodefrag,subvol=@home 0 0
As you can see, there are not so many subvolumes as created by default by Yast, but it works.
For various values of 'works. My background comes from civil and aviation engineering where the premise is that the environment, the use is not benign, that you assume things that can go wrong will go wrong and things that can't go wrong will also go wrong. Some people call this 'belt and braces' engineering :-0 I don't know where you get the idea that a separate /boot is not *required* any longer. It may not be mandatory in many use-cases but it makes life a lot simpler in a lot more. There's a similar argument bout having a separate /tmp or even a /tmp on a separate spindle. The latter is about performance, the former is about security. The bug that the having /tmp on a different FS from the rootFS was fixed, but that's not to say some errant programmer won't recreate it. Even so, it doesn't take an errant programmer to come up with a runaway process that can consume all the scape on /tmp. And if /tmp is NOT on its own FS that means its going to eat all your main FS and that will soon result in a frozen system and for some an unbootable system. While I like the optimizations that a BtrFs-as-the-whole-FS-tree offers in terms of a the optimizations and packing it can achieve, I want a robust, that is robust in the face of perversity and my own stupidity and carelessness, system. Those are among the reasons I think that a separate /boot and separate /tmp make sense. Like Lynn and John I'm not a saint. I'm not on the road for beatification and I make typing mistakes, even when editing config files! Keeping with the mainstream so that the advice the other 'core experts' here give me can be applied makes life easier. That Debian or Redhat have their own way of doing things, have made different decisions in many ways, is their concern. As people have commented, there are things that are not in BtrFs, not in systemd for the openSuse version so as to ensure better stability and reliability. OpenSuse is not as 'bleeding edge'. I like that. I asked why you chose not to have a separate /boot and the answer seems to be "'Cos Debian didn't do it that way". It makes me wonder why you have chosen to move from Debian to openSuse as you seem intent on rebuilding your openSuse system so that it looks like a Debian system. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 12:26:53 -0500
Anton Aylward
Some people call this 'belt and braces' engineering :-0
I appreciate your intention to help and just wonder if you maybe like Ada language?
Even so, it doesn't take an errant programmer to come up with a runaway process that can consume all the scape on /tmp.
In my case, it's single-user desktop machine and I'm the only programmer who can fill the /tmp. :-)
I asked why you chose not to have a separate /boot and the answer seems to be "'Cos Debian didn't do it that way".
Wrong! It might be that Debian stable still uses /bot, but I run Linux without /boot for quite some time. Moreover, it's, as pointed out by jjd, even the decision of Suse itself, at least for 13.2. ;)
It makes me wonder why you have chosen to move from Debian to openSuse as you seem intent on rebuilding your openSuse system so that it looks like a Debian system.
There are more things relevant for chosing the distro besides disk layout. Zipper is certainly the one. ;) Sincerely, Gour -- As the ignorant perform their duties with attachment to results, the learned may similarly act, but without attachment, for the sake of leading people on the right path. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 12:48 PM, Gour wrote:
Moreover, it's, as pointed out by jjd, even the decision of Suse itself, at least for 13.2. ;)
I think you have another conceptual problem here. The 'default' is minimal. There is nothing forcing you to take what the default offers. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 14:45:06 -0500
Anton Aylward
I think you have another conceptual problem here. The 'default' is minimal. There is nothing forcing you to take what the default offers.
So, why are you not satisfied with what defaut is offering and which should be safe for majority of users and go to the unsafe path not recommended by the distro creators themselves? :-) Sincerely, Gour -- You have a right to perform your prescribed duty, but you are not entitled to the fruits of action. Never consider yourself the cause of the results of your activities, and never be attached to not doing your duty. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 02:49 PM, Gour wrote:
On Sun, 23 Nov 2014 14:45:06 -0500 Anton Aylward
wrote: I think you have another conceptual problem here. The 'default' is minimal. There is nothing forcing you to take what the default offers.
So, why are you not satisfied with what defaut is offering and which should be safe for majority of users and go to the unsafe path not recommended by the distro creators themselves? :-)
That's a loaded argument. You are assuming that using a separate /boot and /tmp is 'unsafe'. That is not the case. The default isn't, in opinion, so much as default as 'entries at the top of the list of alternatives'. The way that pull-down lists work. As I say, there are simple, easy precautions. Not, as you 'putting words in my mouth' makes out, "unsafe". I've never understood why people spend so much energy arguing why they shouldn't take simple and easy low cost precautions when it would be less effort to take those precautions. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 23/11/2014 20:45, Anton Aylward a écrit :
On 11/23/2014 12:48 PM, Gour wrote:
Moreover, it's, as pointed out by jjd, even the decision of Suse itself, at least for 13.2. ;)
I think you have another conceptual problem here. The 'default' is minimal. There is nothing forcing you to take what the default offers.
sure, but it's the best openSUSE gurus can offer... and, by the way if you use several partitions, you have to multiply them by the number of linux distros installed - for me at least two (I nver upgrade in place my main machine) jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 02:57 PM, jdd wrote:
and, by the way if you use several partitions, you have to multiply them by the number of linux distros installed - for me at least two (I nver upgrade in place my main machine)
That's a specious argument. You know full well that one autoyast script is all that's needed. *NIX has long been about ways to avoid repetitious work1 -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 23/11/2014 21:05, Anton Aylward a écrit :
That's a specious argument. You know full well that one autoyast script is all that's needed. *NIX has long been about ways to avoid repetitious work1
? let alone remember what partition do what on the systems that do not run now? I already have more than 17 partitions on 4 disk on my main machine :-) by the way we are drifting far from the inital post, on this machine I have no BTRFS file system.... most problem are between the chair and the keyboard, reason why keeping things (reasonably) simple is usefull. that don't mean your config do not fit *your* needs, but I often prefere default configs well known by everybody than special config I will certainly be the only people able to fix occasionnally jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 05:15 PM, jdd wrote:
? let alone remember what partition do what on the systems that do not run now? I already have more than 17 partitions on 4 disk on my main machine :-)
KISS. Just three partitions & types. Linux Filesystem - for ext2 for /boot Linux Swap Linux LVM All my disks are like that. Except the ones that are only a file system or only LVM. KISS. Or did you mean file system? Use blkid # blkid -o full /dev/dm-10 /dev/dm-10: LABEL="ROOT" UUID="4d0f475c-2833-4858-96ed-f284b634ffb2" UUID_SUB="852090d3-1d2d-4412-baf3-056851d87505" TYPE="btrfs" Which is the same as # blkid -o full /dev/disk/by-label/ROOT Since I name all my file/label systems (and swap) things are quite manageable. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, Nov 23, 2014 at 10:57 PM, jdd
Le 23/11/2014 20:45, Anton Aylward a écrit :
On 11/23/2014 12:48 PM, Gour wrote:
Moreover, it's, as pointed out by jjd, even the decision of Suse itself, at least for 13.2. ;)
I think you have another conceptual problem here. The 'default' is minimal. There is nothing forcing you to take what the default offers.
sure, but it's the best openSUSE gurus can offer...
and, by the way if you use several partitions, you have to multiply them by the number of linux distros installed - for me at least two (I nver upgrade in place my main machine)
Which is the excellent example of use case I tried to explain - clone system, update, if update is OK - just destroy original system; if update does not look good - simply go back and destroy new clone. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 23/11/2014 21:50, Andrei Borzenkov a écrit :
Which is the excellent example of use case I tried to explain - clone system, update, if update is OK - just destroy original system; if update does not look good - simply go back and destroy new clone.
no. for example this prevent me from using some new features (on 13.2, BTRFS is the most obvious). I always use a fresh install. this alows me to get rid of many old unused things. From time to time I recall fre space from old unused installs, but some time I'm glad to go and found old config long forgotten :-) jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 12:48 PM, Gour wrote:
Even so, it doesn't take an errant programmer to come up with a runaway process that can consume all the scape on /tmp.
In my case, it's single-user desktop machine and I'm the only programmer who can fill the /tmp. :-)
Yes but you run code written by other programmers and we've long ago established that some of the contributions are less than perfect! Then there's the matter of configuration. Dare I say it, but some configurations are less than perfectly applicable to some situations and can cause, lets putt it politely, errant behaviour. HO! A lot of the discussion on this list pertains to that. I said that I assume things that can go wrong will go wrong and things that can't go wrong will also go wrong. This list is a testimony to that! Right now my /tmp is lowly filling up because there are quite a number of applications that create temporary files and don't delete then when they exit. I think the Mozilla suite is among them! I am *sure* that Thunderbird does not clear up the attachments I open. Damn FTP-by-mail! The crontab.daily is supposed to clear out /tmp but it doesn't seem to do a good job. We have large disks these days, so creating a 1G /boot on a 1T drive, the drive costing you around $50, isn't going to hurt. Its a simple precaution, like spending $1 for a CO2/smote alarm battery. There are a lot of simple precautions in life, but I've never understood why people spend s much energy arguing why they shouldn't take them. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 23/11/2014 20:58, Anton Aylward a écrit :
lot of simple precautions in life, but I've never understood why people spend s much energy arguing why they shouldn't take them.
because at some level it's more fuss than dealing once in a life with things going bad? jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 23/11/2014 17:29, Anton Aylward a écrit :
I'd be interested in learning why you chose not to have a separate /boot partition.
default install do not make one jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 10:49 AM, Gour wrote:
I hope, from the above it's clear what's wrong?
It looks to me like grub2 doesn't know about your /@/ part. Since the openSuse YAST installer doesn't know about that I'm not surprised. The reso of the opensuse community seems to have got by without the /@/ - the root at a subvolume of itself, so far. I'm still unclear as to why you are so insistent about begin different. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, 23 Nov 2014 11:33:22 -0500
Anton Aylward
It looks to me like grub2 doesn't know about your /@/ part.
Right.
Since the openSuse YAST installer doesn't know about that I'm not surprised.
Well, Yast is not the only tool to manage one's OS. I believe that e.g. Yast can't create full ISP mail server with virtual domains, using MySQL for handlin domains/accounts/aliases, Dovecot with full SASL auth. etc., but I still like that Yast can help with *many* other tasks.
The reso of the opensuse community seems to have got by without the /@/ - the root at a subvolume of itself, so far.
Well, btrfs & snapshotting are new things in Suse. Those things are not present in Debian (Sid) either, but there, at least, one can configure it manually.
I'm still unclear as to why you are so insistent about begin different.
First of all, I'm in the process of moving from Debian to Suse, and one thing I want is to have *identical* setup to which I'm accustomed for some time already. Moreover, when/if you have need to boot with some rescueOS and you mount your OS, it much nicer to see something like this: $ ls -l /mnt/btrfs/ total 0 drwxr-xr-x 1 root root 210 Lis 18 16:21 @/ drwxr-xr-x 1 root root 20 Tra 23 2014 @home/ than your whole OS populated with files. Sincerely, Gour -- A person is considered still further advanced when he regards honest well-wishers, affectionate benefactors, the neutral, mediators, the envious, friends and enemies, the pious and the sinners all with an equal mind. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 11:55 AM, Gour wrote:
I'm still unclear as to why you are so insistent about begin different.
First of all, I'm in the process of moving from Debian to Suse, and one thing I want is to have *identical* setup to which I'm accustomed for some time already.
Which makes me wonder ... If you want to rebuild openSuse so it looks like your old Debian system, why did you move to openSuse at all? When I moved to AIX from Solaris I didn't try rebuilding AIX to look like Solaris; similarly the move from AIX to IRIX to Linux ... Yes different versions of *NIX "cross pollinate" and it helps to know what other systems have. Knowing the AIX Volume Manager let me pick up LVM quickly even though it was subtly different. Knowing Perl from SUN let me pick up the AIX SPSS quickly. And more. But that's quite different from trying to remake openSuse so its - as you say - *identical* to your old Debian system. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/23/2014 11:55 AM, Gour wrote:
Well, btrfs & snapshotting are new things in Suse.
No they are not. -- /"\ \ / ASCII Ribbon Campaign X Against HTML Mail / \ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Tue, Nov 18, 2014 at 4:47 PM, Gour
Hello,
I'm considering to move from Debian (Sid) to openSuSe where my Linux adventure has begun in 1999. :-)
At the moment I use raid-1 setup of 2 X 1TB GPT-partitioned disks with the following layout:
/dev/sd(a,b)1 bios_grub 2MB /dev/sd(a,b)2 swap 8GB /dev/sd(a,b)3 btrfs 923.51GB
/dev/sd(a,b)2 partitions are in raid1 swap volume, while btrfs partitions are also in raid-1 and I use subvolumes for / & /home:
UUID=some-uuid / btrfs defaults,noatime,compress=lzo,autodefrag,subvol=@ 0 0 UUID=some-uuid /home btrfs defaults,noatime,compress=lzo,autodefrag,subvol=@home 0 0
I'd like to replicate this setup with Suse as well, but, if I'm right, the installer in 13.2 can't do it?
Correct.
The other option, which I used when installing Debian, is to install Suse on single disk and then clone it and add to raid-1 array.
Any hint?
Yes, I did it in previous openSUSE releases (not production system though). Downside is that you cannot use YaST partitioner; also not sure about yast-bootloader - it will likely not understand that there are multiple disks to install bootloader on. Although editing /etc/default/grub_installdevice manually could work. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Tue, 18 Nov 2014 17:27:08 +0300
Andrei Borzenkov
Yes, I did it in previous openSUSE releases (not production system though). Downside is that you cannot use YaST partitioner;
I'm thinking about cofiguring one disk and then manually add the 2nd one into mirror. Otoh, I wonder, since it's official decision to use XFS to /home to follow their hint and wait when /home with btrfs will become official practice, possibly with subvolumes. Do you use XFS or what you recommend for /home?
also not sure about yast-bootloader - it will likely not understand that there are multiple disks to install bootloader on. Although editing /etc/default/grub_installdevice manually could work.
Iirc, I had to manually install Grub2 into MBR on Debian as well, so not a big issue. Sincerely, Gour -- A person is said to be elevated in yoga when, having renounced all material desires, he neither acts for sense gratification nor engages in fruitive activities. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/19/2014 12:33 PM, Gour wrote:
Otoh, I wonder, since it's official decision to use XFS to /home to follow their hint and wait when /home with btrfs will become official practice, possibly with subvolumes.
That's what I did. I'm not sure why they did it this way, but they have spent more time thinking abut it than I did, so I went with it. Also I wanted to encrypt home (and another project directory) and Yast didn't offer encryption for Btrfs, even though the documentation said it can be done. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On November 19, 2014 3:52:17 PM EST, John Andersen
On 11/19/2014 12:33 PM, Gour wrote:
Otoh, I wonder, since it's official decision to use XFS to /home to follow their hint and wait when /home with btrfs will become official practice, possibly with subvolumes.
That's what I did. I'm not sure why they did it this way, but they have spent more time thinking abut it than I did, so I went with it.
Also I wanted to encrypt home (and another project directory) and Yast didn't offer encryption for Btrfs, even though the documentation said it can be done.
I've encrypted my home dir with ecryptfs over btrfs and have been running that way for a very long time now. This isn't "available" or "supported" by yast or whatever but is fully supported outside that by the kernel, encryptfs-utils, pam-config, etc. -- Later, Darin -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/19/2014 1:52 PM, Darin wrote:
On November 19, 2014 3:52:17 PM EST, John Andersen
wrote: On 11/19/2014 12:33 PM, Gour wrote:
Otoh, I wonder, since it's official decision to use XFS to /home to follow their hint and wait when /home with btrfs will become official practice, possibly with subvolumes.
That's what I did. I'm not sure why they did it this way, but they have spent more time thinking abut it than I did, so I went with it.
Also I wanted to encrypt home (and another project directory) and Yast didn't offer encryption for Btrfs, even though the documentation said it can be done.
I've encrypted my home dir with ecryptfs over btrfs and have been running that way for a very long time now. This isn't "available" or "supported" by yast or whatever but is fully supported outside that by the kernel, encryptfs-utils, pam-config, etc.
Does Btrfs see that as one big file or what? Does it affect snapshots in any way? -- _____________________________________ ---This space for rent--- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On November 19, 2014 4:55:29 PM EST, John Andersen
On 11/19/2014 1:52 PM, Darin wrote:
On November 19, 2014 3:52:17 PM EST, John Andersen
On 11/19/2014 12:33 PM, Gour wrote:
Otoh, I wonder, since it's official decision to use XFS to /home to follow their hint and wait when /home with btrfs will become official practice, possibly with subvolumes.
That's what I did. I'm not sure why they did it this way, but they have spent more time thinking abut it than I did, so I went with it.
Also I wanted to encrypt home (and another project directory) and Yast didn't offer encryption for Btrfs, even though the documentation said it can be done.
I've encrypted my home dir with ecryptfs over btrfs and have been running that way for a very long time now. This isn't "available" or "supported" by yast or whatever but is fully supported outside that by
wrote: the kernel, encryptfs-utils, pam-config, etc. Does Btrfs see that as one big file or what?
No its layered above the file system so any and every file/Dir passes thru the ecryptfs layer and gets encrypted. It's file system agnostic and since its above the FS doesn't require anything special, unlike dmcrypt or the blob encrypted home Dir which is SUSE's default, if you extend the disk or volume.
Does it affect snapshots in any way?
Not that I'm aware of. -- Later, Darin -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (7)
-
Andrei Borzenkov
-
Anton Aylward
-
arvidjaar@gmail.com
-
Darin
-
Gour
-
jdd
-
John Andersen