Re: Plans for BTRFS in Fedora
I've noticed scary reports regarding fragmentation on btrfs, some fairly recent (within last 6 months). I'm interested in considering btrfs for my next f15 install, but should I be concerned about this issue? Is it expected to be resolved? -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2011-02-26 at 17:33-05 Lyos Gemini Norezel lyos.gemininore...@gmail.com wrote: On 02/23/2011 06:38 PM, James Ralston wrote: Separate LVM logical volumes can help mitigate consumption-based DoS attacks. For example: if /tmp and /var/tmp are separate LVM logical volumes, then a runaway/malicious process cannot fill up the entire filesystem merely by filling up /tmp or /var/tmp. For the sake of brevity... I already understand the encrypted volumes argument... but I still fail to see why /tmp, /var/tmp/, /opt, /usr, etc need to have their own partitions. I mentioned one: any filesystem tree that grants regular users write access should have some way to prevent DoS attacks. Making that subdirectory tree a separate filesystem is one way to do it. Another reason to isolate user-writable subdirectory trees to separate filesystems is to make certain types of security attacks more difficult (by removing the ability of a regular user to create a hard link to a file). The more complex a system is... the more likely it is to fail. Generally speaking, yes. But sometimes the benefits provided by the increased complexity are worth the (negligible in this case, IMHO) increase in risk. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
BTRFS vs LVM for VM storage (was: Re: Plans for BTRFS in Fedora)
On Tue, Feb 22, 2011 at 02:51:50PM -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. Sorry I'm a bit late on this gentle discussion, but I have one question about this: I use LVM to store virtual machines, one VM per LV, and it's very good for that. How is BTRFS's performance when used to store VMs (presumably they are stored as files)? Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://libguestfs.org -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: BTRFS vs LVM for VM storage (was: Re: Plans for BTRFS in Fedora)
On Wed, Mar 2, 2011 at 9:23 AM, Richard W.M. Jones rjo...@redhat.com wrote: On Tue, Feb 22, 2011 at 02:51:50PM -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. Sorry I'm a bit late on this gentle discussion, but I have one question about this: I use LVM to store virtual machines, one VM per LV, and it's very good for that. How is BTRFS's performance when used to store VMs (presumably they are stored as files)? Good, but the problem is the default behavior of virt manager is to use fsync for everything, you have to manually go in and set the Cache to None so it will use O_DIRECT, and then it's just as fast as anything else. Not a big deal if you create everything via the command line, kind of annoying if you do it via the GUI, tho all you have to do is say let me edit the options before starting this vm when you first create it, set the cache type and then do the install and you are good to go. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Sat, 2011-02-26 at 17:33 -0500, Lyos Gemini Norezel wrote: On 02/23/2011 04:37 PM, Kevin Kofler wrote: And I'd like to counter-counter-propose that we just stop using ANY kind of subvolumes or volume management by default and just default to plain old partitions. IMHO, LVM causes more problems than it fixes. Sure, you can easily add storage from another disk, but in exchange there's no straightforward way to resize your partitions, at least none of the common partition editors can do it. There's also a performance penalty. +1 This subvolume nonsense has no real place on any home computer/consumer device. With all due respect, that's the path chosen by certain other Linux distributions (ones where if I install them I have to jump through all kinds of hoops to turn on LVM). That is not the way we should be going. I've made my objections known, added a comment on the wiki discussion for the feature, and will raise an objection at the appropriate time that it is proposed to drop LVM use by default. Until then, I'm done :) Jon. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Mon, Feb 28, 2011 at 1:10 PM, Jon Masters jonat...@jonmasters.org wrote: On Sat, 2011-02-26 at 17:33 -0500, Lyos Gemini Norezel wrote: On 02/23/2011 04:37 PM, Kevin Kofler wrote: And I'd like to counter-counter-propose that we just stop using ANY kind of subvolumes or volume management by default and just default to plain old partitions. IMHO, LVM causes more problems than it fixes. Sure, you can easily add storage from another disk, but in exchange there's no straightforward way to resize your partitions, at least none of the common partition editors can do it. There's also a performance penalty. +1 This subvolume nonsense has no real place on any home computer/consumer device. With all due respect, that's the path chosen by certain other Linux distributions (ones where if I install them I have to jump through all kinds of hoops to turn on LVM). That is not the way we should be going. I've made my objections known, added a comment on the wiki discussion for the feature, and will raise an objection at the appropriate time that it is proposed to drop LVM use by default. Until then, I'm done :) I don't see any comments on the discussion page. And I'm not talking about adding extra hoops, you just have to select the custom partition scheme. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/26/2011 05:33 PM, Lyos Gemini Norezel wrote: This subvolume nonsense has no real place on any home computer/consumer device. ... Having more than 3 partitions on ANY system other than production servers seems foolish at best. To have it as default on a modern operating system is nothing short of insanity. You state these as facts but they are really your opinions, and many reasonable people disagree with you to varying degrees. I for instance like a logical volume facility, even if it is awkward like LVM. I am really looking forward to LVs from BTRFS. Note that when the volume management comes as part of the FS, it stops being an issue: if someone doesn't like volumes, they may decline to use them, that's all. I think your arguments would be more effective if you were more precise and stated them as your opinions: I believe that the subvolumes are an overkill on most non-server systems, for the following reasons: Greetings -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 04:37 PM, Kevin Kofler wrote: And I'd like to counter-counter-propose that we just stop using ANY kind of subvolumes or volume management by default and just default to plain old partitions. IMHO, LVM causes more problems than it fixes. Sure, you can easily add storage from another disk, but in exchange there's no straightforward way to resize your partitions, at least none of the common partition editors can do it. There's also a performance penalty. Kevin Kofler +1 This subvolume nonsense has no real place on any home computer/consumer device. On 02/23/2011 06:38 PM, James Ralston wrote: 1. Separate LVM logical volumes can help mitigate consumption-based DoS attacks. For example: if /tmp and /var/tmp are separate LVM logical volumes, then a runaway/malicious process cannot fill up the entire filesystem merely by filling up /tmp or /var/tmp. Could someone please explain to me what the average home user has for a ridiculous amount of partitions? For the sake of brevity... I already understand the encrypted volumes argument... but I still fail to see why /tmp, /var/tmp/, /opt, /usr, etc need to have their own partitions. The more complex a system is... the more likely it is to fail. Having more than 3 partitions on ANY system other than production servers seems foolish at best. To have it as default on a modern operating system is nothing short of insanity. Lyos Gemini Norezel -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Sat, 2011-02-26 at 17:33 -0500, Lyos Gemini Norezel wrote: Having more than 3 partitions on ANY system other than production servers seems foolish at best. To have it as default on a modern operating system is nothing short of insanity. I'm not sure why your mail is so strident, because we don't default to that. The default Fedora layout is either /boot , swap , / or /boot , swap , / , and /home . Okay, that last one is four, but only if you count swap. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org http://www.happyassassin.net -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Dne 27.2.2011 06:51, Adam Williamson napsal(a): I'm not sure why your mail is so strident, because we don't default to that. The default Fedora layout is either /boot , swap , / or /boot , swap , / , and /home . Okay, that last one is four, but only if you count swap. He confuses mountpoints and partitions and the rant is against numerous tmpfs mounted directories. That's at least how I understood him, not that I would care for his argument. Matěj -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Dne 24.2.2011 20:54, Ric Wheeler napsal(a): Can we have pointers to these crashes or BZ reports please? As Josef has noted, btrfs has been quite stable in our testing and we are certainly going to pursue any reports. Will do ... I am hesitant to do so, because so many of my previous bug reports didn't make much difference, and given fsck core dumping (for the last three years), it seems to me that BTRFS is just nothing I would take seriously. Yet. Just to answer your last question, we do not intend to slow it down. Rather, we hope to speed it up considerably by adding developers, testing and users :) I meant obviously slowing down accepting BTRFS as default filesystem for Fedora and throwing away LVM from default, not development of BTRFS itself. Matěj -- http://www.ceplovi.cz/matej/, Jabber: mceplatceplovi.cz GPG Finger: 89EF 4BC6 288A BF43 1BAB 25C3 E09F EF25 D964 84AC smime.p7s Description: S/MIME Cryptographic Signature -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Thu, Feb 24, 2011 at 7:54 PM, Ric Wheeler rwhee...@redhat.com wrote: On 02/24/2011 08:44 AM, Matej Cepl wrote: Dne 23.2.2011 20:49, Matthew Garrett napsal(a): btrfs does the former without anywhere near as much of the latter. BTRFS so far only makes my kernel panicking as it did anytime I have been trying it since Fedora 9 (yes, I am crazy). This is absolutely not meant as anything personal against Josef (I know very well how incredibly small group are BTRFS developers), but just a bit of suspicion, whether we have fsck now (or we will have fsck soon) really leads so quickly let's make it default. I am quite OK with having crashing and unstable systemd or Gnome 3 (and again, nothing against their developers, this is Rawhide and Fedora, so when my kids are alive despite me using it I am pretty happy), but unstable file system is quite a different matter. Could we slow down a bit, please? Matěj Can we have pointers to these crashes or BZ reports please? As Josef has noted, btrfs has been quite stable in our testing and we are certainly going to pursue any reports. Also note that the btrfs community of developers is not so small these days and rivals (if not surpasses) the size of the team working on ext4. Just to answer your last question, we do not intend to slow it down. Rather, we hope to speed it up considerably by adding developers, testing and users :) I've seen a number of crashes using 2.6.37 on a Dell 6410 using btrfs in a luks encrypted LVM volume. Sometimes its a message in dmesg, other times an out right crash. Each time it happens I submit the kernel oops using abrt, but unlike RHBZ reports you don't get a URL for the report so I have no idea where they get reported to but it might be worthwhile reviewing that information where ever it ends up. Peter -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/25/2011 04:06 AM, Peter Robinson wrote: On Thu, Feb 24, 2011 at 7:54 PM, Ric Wheelerrwhee...@redhat.com wrote: On 02/24/2011 08:44 AM, Matej Cepl wrote: Dne 23.2.2011 20:49, Matthew Garrett napsal(a): btrfs does the former without anywhere near as much of the latter. BTRFS so far only makes my kernel panicking as it did anytime I have been trying it since Fedora 9 (yes, I am crazy). This is absolutely not meant as anything personal against Josef (I know very well how incredibly small group are BTRFS developers), but just a bit of suspicion, whether we have fsck now (or we will have fsck soon) really leads so quickly let's make it default. I am quite OK with having crashing and unstable systemd or Gnome 3 (and again, nothing against their developers, this is Rawhide and Fedora, so when my kids are alive despite me using it I am pretty happy), but unstable file system is quite a different matter. Could we slow down a bit, please? Matěj Can we have pointers to these crashes or BZ reports please? As Josef has noted, btrfs has been quite stable in our testing and we are certainly going to pursue any reports. Also note that the btrfs community of developers is not so small these days and rivals (if not surpasses) the size of the team working on ext4. Just to answer your last question, we do not intend to slow it down. Rather, we hope to speed it up considerably by adding developers, testing and users :) I've seen a number of crashes using 2.6.37 on a Dell 6410 using btrfs in a luks encrypted LVM volume. Sometimes its a message in dmesg, other times an out right crash. Each time it happens I submit the kernel oops using abrt, but unlike RHBZ reports you don't get a URL for the report so I have no idea where they get reported to but it might be worthwhile reviewing that information where ever it ends up. Peter I think that it is probably best to report issues to the linux-btrfs list where the developers are. If you report them via bugzilla, we will see them directly there as well. Seems that we need to figure out where these abrt generated BZ's go, I have not seen them come in via our normal bugzilla reports but might need to figure out how to do specific queries for them. Thanks! Ric -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Fri, Feb 25, 2011 at 1:31 PM, Ric Wheeler rwhee...@redhat.com wrote: On 02/25/2011 04:06 AM, Peter Robinson wrote: On Thu, Feb 24, 2011 at 7:54 PM, Ric Wheelerrwhee...@redhat.com wrote: On 02/24/2011 08:44 AM, Matej Cepl wrote: Dne 23.2.2011 20:49, Matthew Garrett napsal(a): btrfs does the former without anywhere near as much of the latter. BTRFS so far only makes my kernel panicking as it did anytime I have been trying it since Fedora 9 (yes, I am crazy). This is absolutely not meant as anything personal against Josef (I know very well how incredibly small group are BTRFS developers), but just a bit of suspicion, whether we have fsck now (or we will have fsck soon) really leads so quickly let's make it default. I am quite OK with having crashing and unstable systemd or Gnome 3 (and again, nothing against their developers, this is Rawhide and Fedora, so when my kids are alive despite me using it I am pretty happy), but unstable file system is quite a different matter. Could we slow down a bit, please? Matěj Can we have pointers to these crashes or BZ reports please? As Josef has noted, btrfs has been quite stable in our testing and we are certainly going to pursue any reports. Also note that the btrfs community of developers is not so small these days and rivals (if not surpasses) the size of the team working on ext4. Just to answer your last question, we do not intend to slow it down. Rather, we hope to speed it up considerably by adding developers, testing and users :) I've seen a number of crashes using 2.6.37 on a Dell 6410 using btrfs in a luks encrypted LVM volume. Sometimes its a message in dmesg, other times an out right crash. Each time it happens I submit the kernel oops using abrt, but unlike RHBZ reports you don't get a URL for the report so I have no idea where they get reported to but it might be worthwhile reviewing that information where ever it ends up. Peter I think that it is probably best to report issues to the linux-btrfs list where the developers are. If you report them via bugzilla, we will see them directly there as well. Seems that we need to figure out where these abrt generated BZ's go, I have not seen them come in via our normal bugzilla reports but might need to figure out how to do specific queries for them. I think the kernel ones get submitted here http://kerneloops.org/ but if not you'd have to look closer at the abrt-addon-kerneloops for details on where its sent. Peter -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/25/2011 08:52 AM, Peter Robinson wrote: On Fri, Feb 25, 2011 at 1:31 PM, Ric Wheelerrwhee...@redhat.com wrote: On 02/25/2011 04:06 AM, Peter Robinson wrote: On Thu, Feb 24, 2011 at 7:54 PM, Ric Wheelerrwhee...@redhat.comwrote: On 02/24/2011 08:44 AM, Matej Cepl wrote: Dne 23.2.2011 20:49, Matthew Garrett napsal(a): btrfs does the former without anywhere near as much of the latter. BTRFS so far only makes my kernel panicking as it did anytime I have been trying it since Fedora 9 (yes, I am crazy). This is absolutely not meant as anything personal against Josef (I know very well how incredibly small group are BTRFS developers), but just a bit of suspicion, whether we have fsck now (or we will have fsck soon) really leads so quickly let's make it default. I am quite OK with having crashing and unstable systemd or Gnome 3 (and again, nothing against their developers, this is Rawhide and Fedora, so when my kids are alive despite me using it I am pretty happy), but unstable file system is quite a different matter. Could we slow down a bit, please? Matěj Can we have pointers to these crashes or BZ reports please? As Josef has noted, btrfs has been quite stable in our testing and we are certainly going to pursue any reports. Also note that the btrfs community of developers is not so small these days and rivals (if not surpasses) the size of the team working on ext4. Just to answer your last question, we do not intend to slow it down. Rather, we hope to speed it up considerably by adding developers, testing and users :) I've seen a number of crashes using 2.6.37 on a Dell 6410 using btrfs in a luks encrypted LVM volume. Sometimes its a message in dmesg, other times an out right crash. Each time it happens I submit the kernel oops using abrt, but unlike RHBZ reports you don't get a URL for the report so I have no idea where they get reported to but it might be worthwhile reviewing that information where ever it ends up. Peter I think that it is probably best to report issues to the linux-btrfs list where the developers are. If you report them via bugzilla, we will see them directly there as well. Seems that we need to figure out where these abrt generated BZ's go, I have not seen them come in via our normal bugzilla reports but might need to figure out how to do specific queries for them. I think the kernel ones get submitted here http://kerneloops.org/ but if not you'd have to look closer at the abrt-addon-kerneloops for details on where its sent. Peter Not sure who monitors all kernel oops reports, but I personally don't see them. If you have a btrfs issue (or other issue with fedora file systems), feel free to drop me an email to make sure that we know about it so we can have a look at it :) ric -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2/25/11 2:54 AM, Matěj Cepl wrote: Dne 24.2.2011 20:54, Ric Wheeler napsal(a): Can we have pointers to these crashes or BZ reports please? As Josef has noted, btrfs has been quite stable in our testing and we are certainly going to pursue any reports. Will do ... I am hesitant to do so, because so many of my previous bug reports didn't make much difference, Do you have pointers to those? Were they on the kernel.org bugzilla, or the red hat bugzilla, the list, or elsewhere? Thanks, -Eric and given fsck core dumping (for the last three years), it seems to me that BTRFS is just nothing I would take seriously. Yet. Just to answer your last question, we do not intend to slow it down. Rather, we hope to speed it up considerably by adding developers, testing and users :) I meant obviously slowing down accepting BTRFS as default filesystem for Fedora and throwing away LVM from default, not development of BTRFS itself. Matěj -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Thu, Feb 24, 2011 at 02:25:26PM +0100, Lennart Poettering wrote: snapshotted every time we perform a package/admin operation (and perhaps also just on regular intervals for good measure), what would we then gain by adding a read-only rootfs to the mix? Security, robustness: you can be sure that nothing tempers with your basic OS tree and it is always in a defined state, unless put in a specific admin mode, where the image may be changed/administered, i.e. / is remounted rw. It'd be nice to support a separate /usr in this case as well, because changes to /etc are usually a different use-case than changes to /usr -- the former is administrator configuration actions, and the latter almost exclusively package updates, installations, or removals. (Installing packages may or may not also entail changes to /etc, of course.) -- Matthew Miller mat...@mattdm.org Senior Systems Architect -- Instructional Research Computing Services Harvard School of Engineering Applied Sciences -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Dne 23.2.2011 20:49, Matthew Garrett napsal(a): btrfs does the former without anywhere near as much of the latter. BTRFS so far only makes my kernel panicking as it did anytime I have been trying it since Fedora 9 (yes, I am crazy). This is absolutely not meant as anything personal against Josef (I know very well how incredibly small group are BTRFS developers), but just a bit of suspicion, whether we have fsck now (or we will have fsck soon) really leads so quickly let's make it default. I am quite OK with having crashing and unstable systemd or Gnome 3 (and again, nothing against their developers, this is Rawhide and Fedora, so when my kids are alive despite me using it I am pretty happy), but unstable file system is quite a different matter. Could we slow down a bit, please? Matěj -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 06:01 PM, James Ralston wrote: On 2011-02-23 at 13:41-05 Peter Jonespjo...@redhat.com wrote: dm-crypt still just throws REQ_FLUSH away instead of figuring out the block remaps involved and issuing the right bios. Of course, this is a problem with dm-crypt and _any_ filesystem. Are you sure that's still the case? Well, when I said So, don't hold me to this, but it looks as if... in the sentence before the one you quoted, I was attempting to imply that I was not, in fact, sure, and had only done the most trivial read of the code in question. Because this patchset appears to add REQ_FLUSH support to dm: http://www.redhat.com/archives/dm-devel/2010-August/msg00303.html ...and that patchset is included in 2.6.38-rc6. The second patch in that patchset is the one I referenced in the email you're replying to. But I guess I misread it and it does seem to say it fixes dm-crypt as well. -- Peter The Shuttle is now going five times the sound of speed. -- Dan Rather, first landing of Columbia -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2011-02-23 at 23:32-06 Michael Cronenworth m...@cchtml.com wrote: On 02/23/2011 05:38 PM, James Ralston wrote: None of these issues is a dealbreaker, but they *are* losses of functionality versus what LVM offers. LVM isn't going anywhere. It just won't be the default during a fresh installation, which you would still be free to override by using an LVM again if you wished. But I *don't* wish to discard btrfs and continue to use LVM. What I want to do is *replace* my use of LVM with btrfs. But to do that, I need to either figure out how to make btrfs do the things I currently do with LVM, or I need to figure out how to change the things I do to better leverage the capabilities of btrfs. I'm open to the former, the latter, or some combination of both. But right now, I've been unable to find enough information about btrfs to do either. A third thing that's important to me (which others have already mentioned): disks for virtual guests. With LVM, I simply lvcreate a new volume, and then when I create a new guest with virt-manager, I tell it to use the block device for the LVM logical volume. This avoids having a filesystem layer between the virtual guest and the LVM block device, which I've assumed (but have not actually tested) is a performance win. But in the btrfs is your entire disk model, it seems that my only option is to create a big file on the filesystem, and tell virt-manager to use that as the guest's disk. Even if I create the big file efficiently (i.e., with fallocate(2)), how much of a performance hit will I take by having the guest's disk I/O go through the btrfs filesystem layer, versus the LVM layer? A little? A lot? None at all, because it'll actually be a performance win? I have no idea. And furthermore, I don't know if there's a better way to do it. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/24/2011 08:44 AM, Matej Cepl wrote: Dne 23.2.2011 20:49, Matthew Garrett napsal(a): btrfs does the former without anywhere near as much of the latter. BTRFS so far only makes my kernel panicking as it did anytime I have been trying it since Fedora 9 (yes, I am crazy). This is absolutely not meant as anything personal against Josef (I know very well how incredibly small group are BTRFS developers), but just a bit of suspicion, whether we have fsck now (or we will have fsck soon) really leads so quickly let's make it default. I am quite OK with having crashing and unstable systemd or Gnome 3 (and again, nothing against their developers, this is Rawhide and Fedora, so when my kids are alive despite me using it I am pretty happy), but unstable file system is quite a different matter. Could we slow down a bit, please? Matěj Can we have pointers to these crashes or BZ reports please? As Josef has noted, btrfs has been quite stable in our testing and we are certainly going to pursue any reports. Also note that the btrfs community of developers is not so small these days and rivals (if not surpasses) the size of the team working on ext4. Just to answer your last question, we do not intend to slow it down. Rather, we hope to speed it up considerably by adding developers, testing and users :) Ric -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Thu, Feb 24, 2011 at 8:44 AM, Matej Cepl mc...@redhat.com wrote: Dne 23.2.2011 20:49, Matthew Garrett napsal(a): btrfs does the former without anywhere near as much of the latter. BTRFS so far only makes my kernel panicking as it did anytime I have been trying it since Fedora 9 (yes, I am crazy). This is absolutely not meant as anything personal against Josef (I know very well how incredibly small group are BTRFS developers), but just a bit of suspicion, whether we have fsck now (or we will have fsck soon) really leads so quickly let's make it default. I am quite OK with having crashing and unstable systemd or Gnome 3 (and again, nothing against their developers, this is Rawhide and Fedora, so when my kids are alive despite me using it I am pretty happy), but unstable file system is quite a different matter. Could we slow down a bit, please? Sure but where's the fun in that? :). Seriously though, when I'm not on RHEL duty most of my time has been spent on stability vs. adding features, and honestly there aren't a lot of OMFG this shit is broken complaints, it's more of hey when I do this thing that is specific to my work load, BTRFS does X wrong. At this point what is left is the normal day to day users using this thing and breaking it in ways that we as developers have not been able to break it using it the way we use it and test it the way we test it. I think that if I could get a large base to test for F15 that we could squash most/all of the problems that crop up from that to be in great shape for default in F16. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2011-02-24 at 16:02-05 Josef Bacik jo...@toxicpanda.com wrote: I think that if I could get a large base to test for F15 that we could squash most/all of the problems that crop up from that to be in great shape for default in F16. I think you'd increase your chances of getting lots of testers for F15+btrfs if there were documentation that explained how to take certain LVM configurations/actions and translate them into btrfs best practices. For examples, see my other follow-ups in this thread... -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 01:26 AM, Josef Bacik wrote: Various things, better data integrity to start with, and if you install the yum-fs-snapshot you have the ability to rollback easily. So we got the above + What Lennart mentioned as benefits to the end user. Now if we continue to hang on to the outdated concept ( From my perspective ) of defaults as opposed to allowing each SIG to control and set it's own defaults that suits their target end user base. For example any of the *DE might not want to switch to BTRFS until all GUI tool for their end user to take advantage of what it has to offer are in place while the Server SIG might certainly do. Hence I propose that we open a ticket with FESCO which will request them to do something like this. Review the current filesystem default Gather feed back from various SIG with in the community to create a good criteria for what needs to be in place in a filesystem for it to be considered as an a default for Fedora. Gather a list of all existing linux filesystems ( there seems to be a new one popping up every other month or so ). Filter out those that do not live up to that criteria and any base line they set for their decision like must not perform worse then the current default must work well on SSD's etc.. settle on one that benefits most likely most the novice end users. ( We experienced users are able to choose tweak and turn any knob at install time to suit our need which makes the novice end user the lowest dominator) JBG -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, Feb 22, 2011 at 10:25 PM, Jon Masters jonat...@jonmasters.org wrote: On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. In my personal opinion, this is a poor design decision. Yes, BTRFS can do a lot of volume-y things, and these are growing by the day, but I don't want my filesystem replacing a full volume manager and I am concerned that this will lead to less testing and exposure to full LVM use within the Fedora community. Instead, I'd like to counter-propose that everything stay exactly as it is, with users being able to elect to switch to BTRFS (sub)volumes if they are interested in doing so. Should the switch to BTRFS by default happen, this will be one more thing I will have to fix immediately during installation. The list grows longer and longer over time - please don't make this change. You seem to spend a lot of time during your installs undoing all the new things that were done for the release. Perhaps a rapid changing, bleeding-edge distribution isn't quite suited to your needs. Maybe you would be more comfortable with Debian or CentOS? josh -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, 22.02.11 22:25, Jon Masters (jonat...@jonmasters.org) wrote: On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. In my personal opinion, this is a poor design decision. Yes, BTRFS can do a lot of volume-y things, and these are growing by the day, but I don't want my filesystem replacing a full volume manager and I am concerned that this will lead to less testing and exposure to full LVM use within the Fedora community. Instead, I'd like to counter-propose that everything stay exactly as it is, with users being able to elect to switch to BTRFS (sub)volumes if they are interested in doing so. Should the switch to BTRFS by default happen, this will be one more thing I will have to fix immediately during installation. The list grows longer and longer over time - please don't make this change. Aren't you exaggerating your conservatism a bit? Are there actually new Fedora features you do support? The only signal you appear to be sending all the time is NO!. Lennart -- Lennart Poettering - Red Hat, Inc. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, Feb 22, 2011 at 10:25 PM, Jon Masters jonat...@jonmasters.org wrote: On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. In my personal opinion, this is a poor design decision. Yes, BTRFS can do a lot of volume-y things, and these are growing by the day, but I don't want my filesystem replacing a full volume manager and I am concerned that this will lead to less testing and exposure to full LVM use within the Fedora community. Instead, I'd like to counter-propose that everything stay exactly as it is, with users being able to elect to switch to BTRFS (sub)volumes if they are interested in doing so. Should the switch to BTRFS by default happen, this will be one more thing I will have to fix immediately during installation. The list grows longer and longer over time - please don't make this change. Theres no sense in using LVM if BTRFS can do the same job better. I'm not suggesting removing LVM altogether, just not using it for the default pick the layout for me option, you can always do your own thing with a custom layout. Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, Feb 22, 2011 at 11:57 PM, Bruno Wolff III br...@wolff.to wrote: On Tue, Feb 22, 2011 at 14:51:50 -0500, Josef Bacik jo...@toxicpanda.com wrote: 3) All the various little tools that we have for putting together LiveCD's that are very ext* centered. I've not even looked at this yet, but I assume it's going to be kind of a pain. I like to see live CDs just use squashfs directly and not a squashed copy of an ext3/4 image. However this will break using dd to copy the image for installs which will slow things down. Further however, there is at least some interest in having live CDs and ananconda share the same install method, so this might be happening in any case. This would be a great thing in general since the default ext* image is shrunk down to be installed which creates a bad fs layout which has performance implications. Sure it may be faster for the install, but overall it hurts our users. So moving away from this dd an image thing would be good for everybody. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 4:25 AM, Jon Masters jonat...@jonmasters.org wrote: On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. In my personal opinion, this is a poor design decision. Yes, BTRFS can do a lot of volume-y things, and these are growing by the day, but I don't want my filesystem replacing a full volume manager and I am concerned that this will lead to less testing and exposure to full LVM use within the Fedora community. Instead, I'd like to counter-propose that everything stay exactly as it is, with users being able to elect to switch to BTRFS (sub)volumes if they are interested in doing so. Should the switch to BTRFS by default happen, this will be one more thing I will have to fix immediately during installation. The list grows longer and longer over time - please don't make this change. Oh god changes how dare we even think about those evil things! Now seriously we cannot make progress when every time a (bigger) change is proposed people start screaming but this changes what we had before, and regarding this particular case I think LVM by default has pretty much always been a bad idea anyway. And it is not like installing an OS is something you do everyday so you install your OS, do the changes you think are necessary and be done with it. Defaults should be chooses on the metric what provides the best experience for the users not based on what we have been doing in the past (i.e stagnation). -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Hi I wanted to second these questions... 2011/2/22 Jóhann B. johan...@gmail.com: Will there be any performance penalties making this move? [...] What benefit will this switch bring to the novice desktop end users? Will the novice desktop end user ever take advantages of any of the features that btrfs brings? Since upgrading (downgrading?) my netbook to use an SSD I went with the Anaconda defaults (using LVM etc) and that probably wasn't in my best interest - judging from benchmarks at the time of read performance in the LVM compared to the underlying device, also from the point of not being able to add complex disk arrays to the netbook any time soon. I think Fedora could do more to support lower end devices *well*, in addition to allowing people to use the very latest technology on larger (ie. desktop and server) platforms. My impression right now is I'd be interested to try BTRFS for the server and maybe larger desktops, but would probably want to avoid it for anything smaller. -Cam -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 8:29 AM, Camilo Mesias cam...@mesias.co.uk wrote: Hi I wanted to second these questions... 2011/2/22 Jóhann B. johan...@gmail.com: Will there be any performance penalties making this move? [...] What benefit will this switch bring to the novice desktop end users? Will the novice desktop end user ever take advantages of any of the features that btrfs brings? Since upgrading (downgrading?) my netbook to use an SSD I went with the Anaconda defaults (using LVM etc) and that probably wasn't in my best interest - judging from benchmarks at the time of read performance in the LVM compared to the underlying device, also from the point of not being able to add complex disk arrays to the netbook any time soon. I think Fedora could do more to support lower end devices *well*, in addition to allowing people to use the very latest technology on larger (ie. desktop and server) platforms. My impression right now is I'd be interested to try BTRFS for the server and maybe larger desktops, but would probably want to avoid it for anything smaller. Your impression is wrong, there has been quite a bit of work done to make BTRFS work well on small devices, it is the default filesystem for meego which goes on phones, which is much smaller than anything you are going to have on your netbook. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Josef, On Wed, Feb 23, 2011 at 1:42 PM, Josef Bacik jo...@toxicpanda.com wrote: Your impression is wrong, there has been quite a bit of work done to make BTRFS work well on small devices, it is the default filesystem for meego which goes on phones, which is much smaller than anything you are going to have on your netbook. Thanks, thanks for the info, I will be sure to test it when it is available. My impression was based on a quick read of the BTRFS FAQ. I saw more related to problems of limited disk space than to SSD support (but granted, what mention there is sounds promising). I will test. -Cam -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/22/2011 10:25 PM, Jon Masters wrote: On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. In my personal opinion, this is a poor design decision. Yes, BTRFS can do a lot of volume-y things, and these are growing by the day, but I don't want my filesystem replacing a full volume manager and I am concerned that this will lead to less testing and exposure to full LVM use within the Fedora community. Instead, I'd like to counter-propose that everything stay exactly as it is, with users being able to elect to switch to BTRFS (sub)volumes if they are interested in doing so. I think this is more than a little bit over the top, but there are some legitimate concerns we do need to keep track of and figure out solutions to. Off the top of my head, we need to figure out what we're doing with encrypted volumes: 1) can btrfs do encrypted volumes? Or, somewhat weirder but possibly better, encrypted volumes with non-encrypted subvolumes? a) if encrypted volumes but not plaintext subvolumes, then we need /boot split out to a separate physical partition, maybe b) if it can do encrypted subvolumes of non-encrypted volumes, we could make a /boot and a /sysroot and mount /sysroot as / I guess c) if no encrypted volumes, or if that functionality won't land at the same time as the basic parts, then we need to talk about UI for disk encryption: I) keeping lvm when encrypt data is checked in the anaconda UI, so we can continue to use luks on a VG in those cases. II) it may be worth investigating identifying machine classes during installation and varying the defaults based on them. We talk about doing this every 3 or 4 releases for various reasons and usually decide it's a terrible idea and do something else. Should the switch to BTRFS by default happen, this will be one more thing I will have to fix immediately during installation. The list grows longer and longer over time - please don't make this change. I think you're confusing necessity and desire here. -- Peter Computers don't make errors. What they do, they do on purpose. -- Dale -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 05:07 AM, drago01 wrote: Defaults should be chooses on the metric what provides the best experience for the users not based on what we have been doing in the past (i.e stagnation). *One* data corruption constitutes EPIC FAIL. Btrfs is too young, and will be for yet a while longer. -- -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 9:18 AM, John Reiser jrei...@bitwagon.com wrote: On 02/23/2011 05:07 AM, drago01 wrote: Defaults should be chooses on the metric what provides the best experience for the users not based on what we have been doing in the past (i.e stagnation). *One* data corruption constitutes EPIC FAIL. Btrfs is too young, and will be for yet a while longer. Well if data corruption is the test then we shouldn't be using Ext4 either, there was one fixed as recently as the beginning of this month. File systems are software like anything else, there will be bugs. Off the top of my head I can think of 3 data corrupters we've had in 4 years of working on BTRFS, and they've all been hard to hit and have not to my knowledge been seen by users, only us developers in testing. BTRFS is young, but we have to start somewhere. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, 2011-02-23 at 09:27 -0500, Josef Bacik wrote: On Wed, Feb 23, 2011 at 9:18 AM, John Reiser jrei...@bitwagon.com wrote: On 02/23/2011 05:07 AM, drago01 wrote: Defaults should be chooses on the metric what provides the best experience for the users not based on what we have been doing in the past (i.e stagnation). *One* data corruption constitutes EPIC FAIL. Btrfs is too young, and will be for yet a while longer. Well if data corruption is the test then we shouldn't be using Ext4 either, there was one fixed as recently as the beginning of this month. File systems are software like anything else, there will be bugs. Off the top of my head I can think of 3 data corrupters we've had in 4 years of working on BTRFS, and they've all been hard to hit and have not to my knowledge been seen by users, only us developers in testing. BTRFS is young, but we have to start somewhere. Thanks, From a user's perspective, I've been using btrfs for about 1.5 years on multiple computers and I've been very happy (particularly on my netbook where the transparent compression increases the disk writes considerably). I had one small problem where btrfs wouldn't mount, but by booting off of a newer kernel I had no problems. Thanks for your hard work on btrfs everyone! Nathaniel -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 03:27 PM, Josef Bacik wrote: On Wed, Feb 23, 2011 at 9:18 AM, John Reiserjrei...@bitwagon.com wrote: On 02/23/2011 05:07 AM, drago01 wrote: Defaults should be chooses on the metric what provides the best experience for the users not based on what we have been doing in the past (i.e stagnation). *One* data corruption constitutes EPIC FAIL. Btrfs is too young, and will be for yet a while longer. Well if data corruption is the test then we shouldn't be using Ext4 either, there was one fixed as recently as the beginning of this month. File systems are software like anything else, there will be bugs. Off the top of my head I can think of 3 data corrupters we've had in 4 years of working on BTRFS, and they've all been hard to hit and have not to my knowledge been seen by users, only us developers in testing. BTRFS is young, but we have to start somewhere. Thanks, I'm actually not that worried about corruption as that is something that can be fixed once discovered. What creeps me out about btrfs at the moment is this: https://btrfs.wiki.kernel.org/index.php/FAQ#Help.21__Btrfs_claims_I.27m_out_of_space.2C_but_it_looks_like_I_should_have_lots_left.21 The fact that the FS needs manual rebalance operations and that these can take a while (even tough this can be done online) doesn't exactly make btrfs the ideal candidate for an end-user desktop system that should pretty much be able to look after itself. I'm actually quite interested in btrfs especially for servers because of it's features but this problem really worries me. Regards, Dennis -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2/23/11 5:38 AM, Jóhann B. Guðmundsson wrote: On 02/23/2011 01:26 AM, Josef Bacik wrote: Various things, better data integrity to start with, and if you install the yum-fs-snapshot you have the ability to rollback easily. So we got the above + What Lennart mentioned as benefits to the end user. Now if we continue to hang on to the outdated concept ( From my perspective ) of defaults as opposed to allowing each SIG to control and set it's own defaults that suits their target end user base. For example any of the *DE might not want to switch to BTRFS until all GUI tool for their end user to take advantage of what it has to offer are in place while the Server SIG might certainly do. Hence I propose that we open a ticket with FESCO which will request them to do something like this. Review the current filesystem default Gather feed back from various SIG with in the community to create a good criteria for what needs to be in place in a filesystem for it to be considered as an a default for Fedora. Gather a list of all existing linux filesystems ( there seems to be a new one popping up every other month or so ). Perhaps, but the list of possibly usable defaults for Fedora is not very long. ext3, ext4, xfs, btrfs for now, I'd say. Anything that popped up last month is not going to be ready for primetime for years. That said, I'd love to see it easy to tweak the default from among the supported anaconda filesystems via kickstart or boot-time option, vs. having to go into the detailed custom storage setup screen. ext3 please is a lot simpler request than custom partitioning with volume managers, and today it's all rolled into one... -Eric Filter out those that do not live up to that criteria and any base line they set for their decision like must not perform worse then the current default must work well on SSD's etc.. settle on one that benefits most likely most the novice end users. ( We experienced users are able to choose tweak and turn any knob at install time to suit our need which makes the novice end user the lowest dominator) JBG -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 10:19 AM, Dennis Jacobfeuerborn denni...@conversis.de wrote: On 02/23/2011 03:27 PM, Josef Bacik wrote: On Wed, Feb 23, 2011 at 9:18 AM, John Reiserjrei...@bitwagon.com wrote: On 02/23/2011 05:07 AM, drago01 wrote: Defaults should be chooses on the metric what provides the best experience for the users not based on what we have been doing in the past (i.e stagnation). *One* data corruption constitutes EPIC FAIL. Btrfs is too young, and will be for yet a while longer. Well if data corruption is the test then we shouldn't be using Ext4 either, there was one fixed as recently as the beginning of this month. File systems are software like anything else, there will be bugs. Off the top of my head I can think of 3 data corrupters we've had in 4 years of working on BTRFS, and they've all been hard to hit and have not to my knowledge been seen by users, only us developers in testing. BTRFS is young, but we have to start somewhere. Thanks, I'm actually not that worried about corruption as that is something that can be fixed once discovered. What creeps me out about btrfs at the moment is this: https://btrfs.wiki.kernel.org/index.php/FAQ#Help.21__Btrfs_claims_I.27m_out_of_space.2C_but_it_looks_like_I_should_have_lots_left.21 The fact that the FS needs manual rebalance operations and that these can take a while (even tough this can be done online) doesn't exactly make btrfs the ideal candidate for an end-user desktop system that should pretty much be able to look after itself. I'm actually quite interested in btrfs especially for servers because of it's features but this problem really worries me. Yes this is one of the more complicated areas of BTRFS and tends to blow up in our faces a lot. That being said it's only a big deal if you tend to run your filesystem close to full a lot, which most people do not. It is an area that we work very hard to make sure it's not a problem, hopefully we have eliminated all of the big problems and you should really only see ENOSPC when you actually fill up the disk. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 03:33 PM, Josef Bacik wrote: I'm actually not that worried about corruption as that is something that can be fixed once discovered. What creeps me out about btrfs at the moment is this: https://btrfs.wiki.kernel.org/index.php/FAQ#Help.21__Btrfs_claims_I.27m_out_of_space.2C_but_it_looks_like_I_should_have_lots_left.21 The fact that the FS needs manual rebalance operations and that these can take a while (even tough this can be done online) doesn't exactly make btrfs the ideal candidate for an end-user desktop system that should pretty much be able to look after itself. I'm actually quite interested in btrfs especially for servers because of it's features but this problem really worries me. Yes this is one of the more complicated areas of BTRFS and tends to blow up in our faces a lot. That being said it's only a big deal if you tend to run your filesystem close to full a lot, which most people do not. It is an area that we work very hard to make sure it's not a problem, hopefully we have eliminated all of the big problems and you should really only see ENOSPC when you actually fill up the disk. Thanks, I'm pretty sure for novice end user desktops the same procedure that would clean up old snapshot on regular intervals could make rebalance operation run at the same time since it can be done online... JBG -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, 23 Feb 2011 10:33:26 -0500 Josef Bacik jo...@toxicpanda.com wrote: On Wed, Feb 23, 2011 at 10:19 AM, Dennis Jacobfeuerborn denni...@conversis.de wrote: On 02/23/2011 03:27 PM, Josef Bacik wrote: On Wed, Feb 23, 2011 at 9:18 AM, John Reiserjrei...@bitwagon.com wrote: On 02/23/2011 05:07 AM, drago01 wrote: Defaults should be chooses on the metric what provides the best experience for the users not based on what we have been doing in the past (i.e stagnation). *One* data corruption constitutes EPIC FAIL. Btrfs is too young, and will be for yet a while longer. Well if data corruption is the test then we shouldn't be using Ext4 either, there was one fixed as recently as the beginning of this month. File systems are software like anything else, there will be bugs. Off the top of my head I can think of 3 data corrupters we've had in 4 years of working on BTRFS, and they've all been hard to hit and have not to my knowledge been seen by users, only us developers in testing. BTRFS is young, but we have to start somewhere. Thanks, I'm actually not that worried about corruption as that is something that can be fixed once discovered. What creeps me out about btrfs at the moment is this: https://btrfs.wiki.kernel.org/index.php/FAQ#Help.21__Btrfs_claims_I.27m_out_of_space.2C_but_it_looks_like_I_should_have_lots_left.21 The fact that the FS needs manual rebalance operations and that these can take a while (even tough this can be done online) doesn't exactly make btrfs the ideal candidate for an end-user desktop system that should pretty much be able to look after itself. I'm actually quite interested in btrfs especially for servers because of it's features but this problem really worries me. Yes this is one of the more complicated areas of BTRFS and tends to blow up in our faces a lot. That being said it's only a big deal if you tend to run your filesystem close to full a lot, which most people do not. It is an area that we work very hard to make sure it's not a problem, hopefully we have eliminated all of the big problems and you should really only see ENOSPC when you actually fill up the disk. Thanks, Sorry josef, but I do that all the time with my Virtual Machines, as I do not give them more space then strictly needed. I did that to the point that I needed to uninstall a few devel packages in order to upgrade from f14 to rawhide on a VM ... I am, not sure how common it is on a desktop, just wanted to point out it is a use case to be able to run with little disk at least for development VMs. (I guess I can manually run whatever tool there, as long as it is clearly recognizable when I need to do so). Simo. -- Simo Sorce * Red Hat, Inc * New York -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, 2011-02-23 at 07:15 -0500, Josh Boyer wrote: On Tue, Feb 22, 2011 at 10:25 PM, Jon Masters jonat...@jonmasters.org wrote: On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. In my personal opinion, this is a poor design decision. Yes, BTRFS can do a lot of volume-y things, and these are growing by the day, but I don't want my filesystem replacing a full volume manager and I am concerned that this will lead to less testing and exposure to full LVM use within the Fedora community. Instead, I'd like to counter-propose that everything stay exactly as it is, with users being able to elect to switch to BTRFS (sub)volumes if they are interested in doing so. Should the switch to BTRFS by default happen, this will be one more thing I will have to fix immediately during installation. The list grows longer and longer over time - please don't make this change. You seem to spend a lot of time during your installs undoing all the new things that were done for the release. Perhaps a rapid changing, bleeding-edge distribution isn't quite suited to your needs. Maybe you would be more comfortable with Debian or CentOS? I'll bite. I am, indeed, a fan of various Enterprise Linux distributions and I make no pretense that I am not. I do run an Enterprise Linux on one desktop, and it's also true that I intentionally run my primary desktop several Fedora releases behind so as to avoid many of the problems I see from some of these changes. However, I also run more recent Fedora releases, and on those releases, I typically have to undo changes such as the one originally being proposed (replacing LVM). Again, I feel the solution is to have a Fedora architect whose role is to realize the problems caused by seemingly isolated changes, and stop them from propagating. You don't just replace years of UNIX (or Linux) history/heritage overnight without bothering a chunk of the users. Jon. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, 23.02.11 11:41, Jon Masters (jonat...@jonmasters.org) wrote: You seem to spend a lot of time during your installs undoing all the new things that were done for the release. Perhaps a rapid changing, bleeding-edge distribution isn't quite suited to your needs. Maybe you would be more comfortable with Debian or CentOS? I'll bite. I am, indeed, a fan of various Enterprise Linux distributions and I make no pretense that I am not. I do run an Enterprise Linux on one desktop, and it's also true that I intentionally run my primary desktop several Fedora releases behind so as to avoid many of the problems I see from some of these changes. However, I also run more recent Fedora releases, and on those releases, I typically have to undo changes such as the one originally being proposed (replacing LVM). Again, I feel the solution is to have a Fedora architect whose role is to realize the problems caused by seemingly isolated changes, and stop them from propagating. You don't just replace years of UNIX (or Linux) history/heritage overnight without bothering a chunk of the users. We have that already. It's called FESCO, and is pluralistic and democratic and stuff. Lennart -- Lennart Poettering - Red Hat, Inc. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 11:41 AM, Jon Masters wrote: Again, I feel the solution is to have a Fedora architect whose role is to realize the problems caused by seemingly isolated changes, and stop them from propagating. Fedora historically relies on an open source model for this - there are a lot of smart people working on Fedora, and we depend on each and every one of them to realize the problems caused by seemingly isolated changes. With that in mind, I'm wondering if you have any input as to what problems there might be with the change in question. You don't just replace years of UNIX (or Linux) history/heritage overnight without bothering a chunk of the users. This is something what we've historically been willing to do in Linux, and so far we've had pretty good results. -- Peter RFC 882 put the dots in .com. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2/23/11 5:00 AM, Josef Bacik wrote: This would be a great thing in general since the default ext* image is shrunk down to be installed which creates a bad fs layout which has performance implications. Can you expand upon this more? The filesystem is shrunk down when the live image is built, but then once it is installed the fs is resized to fill out the physical disk / partition it is being installed to. -- Jesse Keating Fedora -- Freedom² is a feature! identi.ca: http://identi.ca/jkeating -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wednesday 23 February 2011 15:07:55 Peter Jones wrote: 1) can btrfs do encrypted volumes? Not yet. Although this was a planned feature at some point, according to Josef, nobody has done it yet. If you want to stack it on top of dm-crypt there are caveats as well. From btrfs-wiki: btrfs volumes on top of dm-crypt block devices (and possibly LVM) require write-caching to be turned off on the underlying HDD. Failing to do so, in the event of a power failure, may result in corruption not yet handled by btrfs code. (2.6.33) Is this still prevailing as of 2.6.38? While it may work if you're lucky (at least it did for me since F13) it can also munch your data. That's what happened to me last night in F15 (didn't lose any data though as such stuff was somewhat expected). Btrfs root partition (no LVM) on a dm-crypt volume on Intel 32 GB solid state drive. After an unclean shutdown I was no longer able to mount the filesystem and therefore start the system. I even managed to crash a live system when trying to mount the decrypted luks volume containing the btrfs. As I didn't have much time I just dd'ed an image to a spare disk and reinstalled. If there is any debugging value in this, I can make it available as long as I can be sure all data could be stripped from it (btrfs-image does exactly this, right?). Be aware that the crash-on-shutdown didn't happen with an official Fedora kernel. Due to the update embargo on F15 I built an updated Kernel checked out with fedpkg. If btrfs should become the default filesystem for Fedora (which I strongly support) a nice and clean solution for encrypted filesystems has to be found. Falling back to ext4 lvm when the encryption checkbox gets ticked is just plain ugly. Not supporting encrypted disks at all would be even worse. Any ideas? -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2/23/11 11:42 AM, Jesse Keating wrote: On 2/23/11 5:00 AM, Josef Bacik wrote: This would be a great thing in general since the default ext* image is shrunk down to be installed which creates a bad fs layout which has performance implications. Can you expand upon this more? The filesystem is shrunk down when the live image is built, but then once it is installed the fs is resized to fill out the physical disk / partition it is being installed to. Shrinking down the filesystem tends to scramble it a bit. All the allocator decisions that were made based on fs size at the time of install go out the window as resize2fs fills in every hole it can find to maximally compress the fs. I haven't done a real analysis but I would expect that inode::data locality and filesystem fragmentation suffer as a result. Resizing it back up after the install just attaches free space to the end of the shrunken fs. -Eric -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 11:41:49AM -0500, Jon Masters wrote: Again, I feel the solution is to have a Fedora architect whose role is to realize the problems caused by seemingly isolated changes, and stop them from propagating. You don't just replace years of UNIX (or Linux) history/heritage overnight without bothering a chunk of the users. LVM is functional for enterprise environments but awful for the common home or office cases. If btrfs lives up to its promises it'll give us something that provides pretty much all the functional beneft of LVM without the additional abstraction that makes seemingly straightforward tasks sufficiently awkward that even I find it abstruse (and I hack on ACPI). I think that's entirely in keeping with Linux heritage. (AIX, on the other hand, delights in partying like it's 1979. Are you confusing the two?) -- Matthew Garrett | mj...@srcf.ucam.org -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 11:54:58AM -0600, Eric Sandeen wrote: On 2/23/11 11:42 AM, Jesse Keating wrote: On 2/23/11 5:00 AM, Josef Bacik wrote: This would be a great thing in general since the default ext* image is shrunk down to be installed which creates a bad fs layout which has performance implications. Can you expand upon this more? The filesystem is shrunk down when the live image is built, but then once it is installed the fs is resized to fill out the physical disk / partition it is being installed to. Shrinking down the filesystem tends to scramble it a bit. All the allocator decisions that were made based on fs size at the time of install go out the window as resize2fs fills in every hole it can find to maximally compress the fs. Is there some sort of sparsifying defrag we could do after copying the image over? The dd-based install is so fast that we could actually afford another pass if we needed it IMHO. --CJD -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2/23/11 12:15 PM, Casey Dahlin wrote: On Wed, Feb 23, 2011 at 11:54:58AM -0600, Eric Sandeen wrote: On 2/23/11 11:42 AM, Jesse Keating wrote: On 2/23/11 5:00 AM, Josef Bacik wrote: This would be a great thing in general since the default ext* image is shrunk down to be installed which creates a bad fs layout which has performance implications. Can you expand upon this more? The filesystem is shrunk down when the live image is built, but then once it is installed the fs is resized to fill out the physical disk / partition it is being installed to. Shrinking down the filesystem tends to scramble it a bit. All the allocator decisions that were made based on fs size at the time of install go out the window as resize2fs fills in every hole it can find to maximally compress the fs. Is there some sort of sparsifying defrag we could do after copying the image over? The dd-based install is so fast that we could actually afford another pass if we needed it IMHO. Not really... e4defrag could maybe, in theory, but its still pretty unloved. I wonder how it'd go to prep the install on a sparse image of some reasonable size, and then do an efficient sparse copy of that onto the target. That would save the scrambling, at least, and still give us minimum space usage on the install disc. -Eric --CJD -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 01:15 PM, Casey Dahlin wrote: On Wed, Feb 23, 2011 at 11:54:58AM -0600, Eric Sandeen wrote: On 2/23/11 11:42 AM, Jesse Keating wrote: On 2/23/11 5:00 AM, Josef Bacik wrote: This would be a great thing in general since the default ext* image is shrunk down to be installed which creates a bad fs layout which has performance implications. Can you expand upon this more? The filesystem is shrunk down when the live image is built, but then once it is installed the fs is resized to fill out the physical disk / partition it is being installed to. Shrinking down the filesystem tends to scramble it a bit. All the allocator decisions that were made based on fs size at the time of install go out the window as resize2fs fills in every hole it can find to maximally compress the fs. Is there some sort of sparsifying defrag we could do after copying the image over? The dd-based install is so fast that we could actually afford another pass if we needed it IMHO. Just as soon as you write it ;) -- Peter RFC 882 put the dots in .com. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 12:50 PM, Lars Seipel wrote: On Wednesday 23 February 2011 15:07:55 Peter Jones wrote: 1) can btrfs do encrypted volumes? Not yet. Although this was a planned feature at some point, according to Josef, nobody has done it yet. If you want to stack it on top of dm-crypt there are caveats as well. Right, which is what we'd wind up doing in the encrypted case. From btrfs-wiki: btrfs volumes on top of dm-crypt block devices (and possibly LVM) require write-caching to be turned off on the underlying HDD. Failing to do so, in the event of a power failure, may result in corruption not yet handled by btrfs code. (2.6.33) Is this still prevailing as of 2.6.38? So, don't hold me to this, but it looks as if the normal lvm behaviour was fixed at least as of d87f4c14f2 . That being said, dm-crypt still just throws REQ_FLUSH away instead of figuring out the block remaps involved and issuing the right bios. Of course, this is a problem with dm-crypt and _any_ filesystem. -- Peter RFC 882 put the dots in .com. 01234567890123456789012345678901234567890123456789012345678901234567890123456789 -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
BTRFS on servers (was Re: Plans for BTRFS in Fedora)
On Wed, 2011-02-23 at 16:19 +0100, Dennis Jacobfeuerborn wrote: I'm actually quite interested in btrfs especially for servers because of it's features For what it's worth, we've been running btrfs on our school fileservers since September. After a few teething problems (fixed by increasing /proc/sys/vm/min_free_kbytes), we've had pretty much zero trouble. The best part is we no longer need LVM on those servers and we get the benefit of painless snapshot generation (we tried LVM snapshots and they were *slow*). Jonathan signature.asc Description: This is a digitally signed message part -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: BTRFS on servers (was Re: Plans for BTRFS in Fedora)
On Wed, Feb 23, 2011 at 2:00 PM, Jonathan Dieter jdie...@lesbg.com wrote: On Wed, 2011-02-23 at 16:19 +0100, Dennis Jacobfeuerborn wrote: I'm actually quite interested in btrfs especially for servers because of it's features For what it's worth, we've been running btrfs on our school fileservers since September. After a few teething problems (fixed by increasing /proc/sys/vm/min_free_kbytes), we've had pretty much zero trouble. Wait what? I know we use lots o ram, but you shouldn't have to bump min_free_kbytes. What were you seeing? Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Once upon a time, Matthew Garrett mj...@srcf.ucam.org said: LVM is functional for enterprise environments but awful for the common home or office cases. Define awful. I make use of it all the time on home and office desktops and even my notebook computer. It makes it easy to reassign disk space from purpose A to purpose B (it would be easier if there was a way to shrink a mounted ext3/ext4 FS, but that's a hard problem). Logical volumes for virtual machines perform better (and virtualization is something becoming more common on the desktop for compatibility). -- Chris Adams cmad...@hiwaay.net Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: BTRFS on servers (was Re: Plans for BTRFS in Fedora)
On Wed, 2011-02-23 at 14:18 -0500, Josef Bacik wrote: On Wed, Feb 23, 2011 at 2:00 PM, Jonathan Dieter jdie...@lesbg.com wrote: On Wed, 2011-02-23 at 16:19 +0100, Dennis Jacobfeuerborn wrote: I'm actually quite interested in btrfs especially for servers because of it's features For what it's worth, we've been running btrfs on our school fileservers since September. After a few teething problems (fixed by increasing /proc/sys/vm/min_free_kbytes), we've had pretty much zero trouble. Wait what? I know we use lots o ram, but you shouldn't have to bump min_free_kbytes. What were you seeing? Thanks, Sorry, I didn't mean to imply it was btrfs that was causing the problem. We got new servers when we did our upgrade and ran into problems with the e1000e kernel module being unable to allocate memory even though there was loads free. Bumping up min_free_kbytes fixed it. See https://bugzilla.redhat.com/show_bug.cgi?id=626851 for more detail. Jonathan signature.asc Description: This is a digitally signed message part -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 01:38:05PM -0600, Chris Adams wrote: Define awful. I make use of it all the time on home and office desktops and even my notebook computer. It makes it easy to reassign disk space from purpose A to purpose B (it would be easier if there was a way to shrink a mounted ext3/ext4 FS, but that's a hard problem). Logical volumes for virtual machines perform better (and virtualization is something becoming more common on the desktop for compatibility). You can't move PVs. You need a separate /boot. If you use more than one disk then it adds significant fragility to the boot process. It slows down booting. It provides some functionality that's hugely useful in a small number of cases, and in every other case it just makes your life more complicated. btrfs does the former without anywhere near as much of the latter. -- Matthew Garrett | mj...@srcf.ucam.org -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 07:49:49PM +, Matthew Garrett wrote: You can't move PVs. You need a separate /boot. If you use more than one disk then it adds significant fragility to the boot process. It slows down booting. It provides some functionality that's hugely useful in a small number of cases, and in every other case it just makes your life more complicated. btrfs does the former without anywhere near as much of the latter. Also, the management tools are worse than AIX had in 1992. I'll accept that that's not an inherent flaw in the implementation, merely an inconvenient aspect of reality. -- Matthew Garrett | mj...@srcf.ucam.org -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: BTRFS on servers (was Re: Plans for BTRFS in Fedora)
On Wed, Feb 23, 2011 at 2:42 PM, Jonathan Dieter jdie...@lesbg.com wrote: On Wed, 2011-02-23 at 14:18 -0500, Josef Bacik wrote: On Wed, Feb 23, 2011 at 2:00 PM, Jonathan Dieter jdie...@lesbg.com wrote: On Wed, 2011-02-23 at 16:19 +0100, Dennis Jacobfeuerborn wrote: I'm actually quite interested in btrfs especially for servers because of it's features For what it's worth, we've been running btrfs on our school fileservers since September. After a few teething problems (fixed by increasing /proc/sys/vm/min_free_kbytes), we've had pretty much zero trouble. Wait what? I know we use lots o ram, but you shouldn't have to bump min_free_kbytes. What were you seeing? Thanks, Sorry, I didn't mean to imply it was btrfs that was causing the problem. We got new servers when we did our upgrade and ran into problems with the e1000e kernel module being unable to allocate memory even though there was loads free. Bumping up min_free_kbytes fixed it. See https://bugzilla.redhat.com/show_bug.cgi?id=626851 for more detail. Ah ok, good, nevermind then :), Joesf -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Once upon a time, Matthew Garrett mj...@srcf.ucam.org said: You can't move PVs. What do you think pvmove does? You need a separate /boot. That's needed for more than just LVM (and probably won't go away, as it is a lot simpler to handle a single method in the installer). If you use more than one disk then it adds significant fragility to the boot process. How does it do that (any more than any other multi-disk setup)? It slows down booting. Cite numbers? It was slower early on, but it goes right by now. I don't doubt it adds some time (of course), but I don't see it being any significant amount. It provides some functionality that's hugely useful in a small number of cases, and in every other case it just makes your life more complicated. For most, it doesn't make things any more complicated (because they never touch it). I'd say that LVM is useful in a growing number of cases. btrfs does the former without anywhere near as much of the latter. Oh, I don't object to btrfs and having the basic volume management in the filesystem layer. AdvFS on DEC Unix was great in that respect. I object to your painting of LVM aw awful. -- Chris Adams cmad...@hiwaay.net Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Hi. On Wed, 23 Feb 2011 13:38:05 -0600, Chris Adams wrote Define awful. I make use of it all the time on home and office desktops and even my notebook computer. It makes it easy to reassign disk space from purpose A to purpose B (it would be easier if there was a way to shrink a mounted ext3/ext4 FS, but that's a hard problem). Logical volumes for virtual machines perform better (and virtualization is something becoming more common on the desktop for compatibility). If you never tried the kind of freedom BTRFS and ZFS give you for shifting around disk space, try it. Seriously. Then you'll see where the awful comes from. In perspective it really is. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 02:08:08PM -0600, Chris Adams wrote: Once upon a time, Matthew Garrett mj...@srcf.ucam.org said: You can't move PVs. What do you think pvmove does? Move PEs from one PV to another. You can't move a PV. You need a separate /boot. That's needed for more than just LVM (and probably won't go away, as it is a lot simpler to handle a single method in the installer). Unless you're booting off iSCSI or really, really love software RAID (which btrfs also kindly simplifies), the only reason you need a separate /boot is because we're slow at putting updated filesystem support in grub because there's no need because we default to a separate /boot. There's a certain amount of circularity there. If you use more than one disk then it adds significant fragility to the boot process. How does it do that (any more than any other multi-disk setup)? My experience with LVM is that any number of error states dump me into an awkward recovery scenario. Again it's not an inherent problem with the technology, more just that because nobody outside the enterprise world really cares about LVM nobody has bothered making it easy for non-sysadmins to handle this stuff. It slows down booting. Cite numbers? It was slower early on, but it goes right by now. I don't doubt it adds some time (of course), but I don't see it being any significant amount. Talk to Lennart about it. It's on the order of two seconds. btrfs does the former without anywhere near as much of the latter. Oh, I don't object to btrfs and having the basic volume management in the filesystem layer. AdvFS on DEC Unix was great in that respect. I object to your painting of LVM aw awful. My objections to the technology aren't that strong. My objections to the user experience are. -- Matthew Garrett | mj...@srcf.ucam.org -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 03:33 PM, Matthew Garrett wrote: On Wed, Feb 23, 2011 at 02:08:08PM -0600, Chris Adams wrote: Once upon a time, Matthew Garrettmj...@srcf.ucam.org said: You can't move PVs. What do you think pvmove does? Move PEs from one PV to another. You can't move a PV. You need a separate /boot. That's needed for more than just LVM (and probably won't go away, as it is a lot simpler to handle a single method in the installer). Unless you're booting off iSCSI or really, really love software RAID (which btrfs also kindly simplifies), the only reason you need a separate /boot is because we're slow at putting updated filesystem support in grub because there's no need because we default to a separate /boot. There's a certain amount of circularity there. Also we need to have the initramfs on an unencrypted chunk of disk. If you use more than one disk then it adds significant fragility to the boot process. How does it do that (any more than any other multi-disk setup)? My experience with LVM is that any number of error states dump me into an awkward recovery scenario. Again it's not an inherent problem with the technology, more just that because nobody outside the enterprise world really cares about LVM nobody has bothered making it easy for non-sysadmins to handle this stuff. Yep. -- Peter RFC 882 put the dots in .com. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Once upon a time, Matthew Garrett mj...@srcf.ucam.org said: On Wed, Feb 23, 2011 at 02:08:08PM -0600, Chris Adams wrote: Once upon a time, Matthew Garrett mj...@srcf.ucam.org said: You can't move PVs. What do you think pvmove does? Move PEs from one PV to another. You can't move a PV. Not exactly; pvmove moves data from one PE to another (either in another PV or in the same PV). I guess I don't understand your original objection though. You can move data around within a PV (granted the command usage to do this is non-obvious and it would be good for that to improve, but it is also not a common use case) and from one PV to another; what else is there to move? You need a separate /boot. That's needed for more than just LVM (and probably won't go away, as it is a lot simpler to handle a single method in the installer). Unless you're booting off iSCSI or really, really love software RAID (which btrfs also kindly simplifies), Using software RAID means really, really love it? How else do you propose to boot from software RAID (if you just kinda like it)? We're talking about the typical home/office desktop, so you aren't going to have hardware RAID, and Linux software RAID is superior to motherboard software RAID. the only reason you need a separate /boot is because we're slow at putting updated filesystem support in grub because there's no need because we default to a separate /boot. There's a certain amount of circularity there. Encrypted root filesystem requires a separate (non-encrypted) /boot, and I think that's only going to increase in popularity (especially on notebooks and other mobile devices). My experience with LVM is that any number of error states dump me into an awkward recovery scenario. I guess in the years I've used LVM, I haven't run into that any number of error states (at least AFAIK, but since you didn't identify any of them, it is hard to say). Sure, when LVM was new, I had to learn a few more commands to handle the system didn't boot case, but IIRC only once was that related to LVM (and that was because I renamed a VG without rebuilding the initrd, which may be no longer an issue with dracut). If you don't know what you are doing, just fsck is going to be a problem. That's an issue with documentation and maybe making better recovery tools, not LVM. Adding a new set of commands to learn with BTRFS isn't magically going to make it easier; you still have to learn new stuff. -- Chris Adams cmad...@hiwaay.net Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Once upon a time, Ralf Ertzinger fed...@camperquake.de said: If you never tried the kind of freedom BTRFS and ZFS give you for shifting around disk space, try it. Seriously. Then you'll see where the awful comes from. In perspective it really is. You cut out the parts of my email where I said I don't have any problem with BTRFS, and I look forward to it making things better. As I said, I used AdvFS on DEC Unix for many years; ZFS and BTRFS are late-comers to the filesystem/LVM party (although they obviously bring more to the table). -- Chris Adams cmad...@hiwaay.net Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Wed, Feb 23, 2011 at 4:37 PM, Kevin Kofler kevin.kof...@chello.at wrote: Jon Masters wrote: In my personal opinion, this is a poor design decision. Yes, BTRFS can do a lot of volume-y things, and these are growing by the day, but I don't want my filesystem replacing a full volume manager and I am concerned that this will lead to less testing and exposure to full LVM use within the Fedora community. Instead, I'd like to counter-propose that everything stay exactly as it is, with users being able to elect to switch to BTRFS (sub)volumes if they are interested in doing so. And I'd like to counter-counter-propose that we just stop using ANY kind of subvolumes or volume management by default and just default to plain old partitions. IMHO, LVM causes more problems than it fixes. Sure, you can easily add storage from another disk, but in exchange there's no straightforward way to resize your partitions, at least none of the common partition editors can do it. There's also a performance penalty. Sorry I should clarify, when I say use Btrfs's volume management stuff I mean just doing normal partitions and then creating a Btrfs filesystem and then add disks to the fs as required. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 07:41 PM, Peter Jones wrote: On 02/23/2011 12:50 PM, Lars Seipel wrote: If you want to stack it on top of dm-crypt there are caveats as well. Right, which is what we'd wind up doing in the encrypted case. From btrfs-wiki: btrfs volumes on top of dm-crypt block devices (and possibly LVM) require write-caching to be turned off on the underlying HDD. Failing to do so, in the event of a power failure, may result in corruption not yet handled by btrfs code. (2.6.33) This is completely nonsense and FUD. We asked several times for reproducer and nobody from btrfs camp was able to provide one. If you have reproducer, please send it to me and I'll fix dmcrypt code. But I am almost sure there is no such problem in current code. So, don't hold me to this, but it looks as if the normal lvm behaviour was fixed at least as of d87f4c14f2 . That being said, dm-crypt still just throws REQ_FLUSH away instead of figuring out the block remaps involved and issuing the right bios. Of course, this is a problem with dm-crypt and _any_ filesystem. Another nonsense. See 2.6.37 source code, REQ_FLUSH is processed by dm-crypt correctly. It supported even barriers some time before (at least in 2.6.36). Anything else for today? Milan -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 2011-02-22 at 14:51-05 Josef Bacik jo...@toxicpanda.com wrote: Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. I don't think btrfs subvolumes are capable of replacing LVM functionality quite yet. Here are two usage cases that I care about: 1. Separate LVM logical volumes can help mitigate consumption-based DoS attacks. For example: if /tmp and /var/tmp are separate LVM logical volumes, then a runaway/malicious process cannot fill up the entire filesystem merely by filling up /tmp or /var/tmp. In contrast, although the btrfs wiki mentions the ability to set a quota on subvolumes at creation time, I don't think anything in btrfs-progs implements that. Plus, even if there were a way to set it, I don't see that anything reports it. 2. Separate LVM logical volumes permit easier installs. Primarily because I've never trusted the upgrade process, but also for other reasons, I always perform a full install to upgrade from one version of Fedora to the next. With LVM logical volumes, this process is trivial: I do a custom partition layout, and tell anaconda to format the /boot, /, /tmp, /usr, /var, and /var/cache logical volumes, and leave all the other logical volumes in the system (e.g., /home, /var/log) alone. I can't do this with one big btrfs filesystem with subvolumes. What I would have to do is first boot the installer in rescue mode, then run a bunch of find /subvolume -xdev commands to delete all the data on the subvolumes that I wanted to reformat. Then I'd have to re-run the installer and tell it to use the / btrfs filesystem as-is, without formatting. None of these issues is a dealbreaker, but they *are* losses of functionality versus what LVM offers. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/23/2011 01:33 AM, Josef Bacik wrote: Well I don't think cleaning up the existing patches will be that big of a deal, it's mostly a matter of testing. The problem with GRUB2 is it's GPLv3, explicitly to be a giant pain in the ass for porting any new fs to GRUB since we're all GPLv2 only. Aren't open source licenses grand? Refer to http://lwn.net/Articles/429607/ Rahul -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, 22 Feb 2011 14:51:50 -0500 Josef Bacik jo...@toxicpanda.com wrote: Hello, So we're getting close to having a working fsck tool so I wanted to take the opportunity to talk about the future of BTRFS in Fedora. ...snip... 1) GRUB support. Edward Shishkin did GRUB1 patches for BTRFS a while ago, but they were obviously never merged upstream and were also not included into fedora. These would either need to be cleaned up and put into our grub package, or we'd need to put /boot on a different filesystem. I personally hate the idea of having a non-btrfs /boot partition but I'm not the one in charge of GRUB. Perhaps if we are going to the pain of getting btrfs patches working and stable, we could just look at moving to grub2? ...snip... So what are your thoughts? Thanks, Sounds good. I'd suggest filing a feature page already and getting it setup with all the status, etc. That can be a central point for coordinating that effort. Josef kevin signature.asc Description: PGP signature -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
1) Fedora 16 ships with BTRFS as the default root filesystem. 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. 2) Anaconda support. I've already talked with Will Woods about this some. Really anaconda will format a normal disk with BTRFS with no problem today, the biggest issue here is adding the volume management stuff and allowing users to create subvolumes via anaconda. Given the slate of changes we have lined up for F16 anaconda (https://fedoraproject.org/wiki/Anaconda/Features), I don't know that adding another major storage change is going to happen. I think in large part, we were hoping to be done with major storage changes for a release or two and work on other less touched areas instead (like the UI). But, perhaps we can find some time somewhere. Knowing what the scope of changes looks like might help with the planning a bit. - Chris -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, Feb 22, 2011 at 3:00 PM, Kevin Fenzi ke...@scrye.com wrote: On Tue, 22 Feb 2011 14:51:50 -0500 Josef Bacik jo...@toxicpanda.com wrote: Hello, So we're getting close to having a working fsck tool so I wanted to take the opportunity to talk about the future of BTRFS in Fedora. ...snip... 1) GRUB support. Edward Shishkin did GRUB1 patches for BTRFS a while ago, but they were obviously never merged upstream and were also not included into fedora. These would either need to be cleaned up and put into our grub package, or we'd need to put /boot on a different filesystem. I personally hate the idea of having a non-btrfs /boot partition but I'm not the one in charge of GRUB. Perhaps if we are going to the pain of getting btrfs patches working and stable, we could just look at moving to grub2? Well I don't think cleaning up the existing patches will be that big of a deal, it's mostly a matter of testing. The problem with GRUB2 is it's GPLv3, explicitly to be a giant pain in the ass for porting any new fs to GRUB since we're all GPLv2 only. Aren't open source licenses grand? ...snip... So what are your thoughts? Thanks, Sounds good. I'd suggest filing a feature page already and getting it setup with all the status, etc. That can be a central point for coordinating that effort. Sounds good, I'll do that, thank you, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
1) GRUB support. Edward Shishkin did GRUB1 patches for BTRFS a while ago, but they were obviously never merged upstream and were also not included into fedora. These would either need to be cleaned up and put into our grub package, or we'd need to put /boot on a different filesystem. I personally hate the idea of having a non-btrfs /boot partition but I'm not the one in charge of GRUB. Perhaps if we are going to the pain of getting btrfs patches working and stable, we could just look at moving to grub2? It's planned: https://fedoraproject.org/wiki/Anaconda/Features/Grub2Migration - Chris -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On 02/22/2011 02:51 PM, Josef Bacik wrote: Hello, So we're getting close to having a working fsck tool so I wanted to take the opportunity to talk about the future of BTRFS in Fedora. Coming up in F15 we're going to have the first release of Fedora where ... So what are your thoughts? Thanks, Josef Exciting indeed - also do you happen to know the status of RAID 5 support ? If I recall correctly, the kernel part was completed some time back .. (several months) but the higher level stuff was not ready back then. thanks gene -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, 22.02.11 14:51, Josef Bacik (jo...@toxicpanda.com) wrote: Hello, So we're getting close to having a working fsck tool so I wanted to take the opportunity to talk about the future of BTRFS in Fedora. Coming up in F15 we're going to have the first release of Fedora where we don't need the special boot option to have the ability to format you filesystem as BTRFS. This is in hopes that we can open it up for wider testing before possibly making it the default filesystem. I realize we're in the early stages of F15, but since filesystems are big and important I'd like to get an idea of the amount of work that needs to still be done to get BTRFS in shape for being Fedora's default filesystem. So here are my goals 1) Fedora 16 ships with BTRFS as the default root filesystem. Hmm, what are your plans regarding hierarchy layout for this? My hope is that one day we can ship a read-only root dir by default, or more specifically a btrfs file system with three subvolumes in it: one read-only one mounted to /, and two writable ones mounted to /home and /var, with /tmp mounted from tmpfs. We came a long way with supporting read-only root dirs and it should mostly work now. In F15 for example /etc/mtab is a symlink, and even smaller stuff like /etc/nologin got moved to /var/run, to make write accesses to /etc unnecessary during normal operation. /etc/resolv.conf is the only thing often updated that's left, but with NM using dnsmasq it's static too. By using btrfs subvolumes doing this kind of seperation into writable and non-writable subtrees we don't have to think anymore about the sizes for those file systems at install time, since they all live in the same fs. If this is adopted package managers and system configuration UIs would need to invoke mount / -o rw,remount before doing their work. So, I'd like to see this implemented one day, maybe the adoption of btrfs is the right time to push this through too? I have not filed a feature page for this, as I am not sure I want to push this on F16, and I don't even know if people in general are onboard with this idea. The benefits of this are mostly security and robustness since we know that the actual subvolume the OS is booted from is always in a consistent state during operation and cannot normally be changed. And of course, we get all kind of magic by doing this because we can easily use snapshots to roll back system upgrades while leaving /home completely untouched. Sorry for hijacking your thread like this, Lennart -- Lennart Poettering - Red Hat, Inc. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
[...btrfs and read-only root,/etc...] Music to my ears. Glad you're working on this. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
Josef Bacik writes: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. Just to clarify -- F16 will still have LVM, but not used by default. I believe that there's a huge number of existing systems that are partitioned using LVM. pgpY1YS4gcSHd.pgp Description: PGP signature -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: So what are your thoughts? Thanks, Will there be any performance penalties making this move? I ran it on my workstation when F13 came out with updates+snapshots and while updating my desktop responsiveness was well not so good. Novice end users tools ( GUI ) and policy to clean those snapshots and perhaps storing those snapshots in it's own directory and or partition rather than directly on / need be in place before we implement this as an default. I suggest people try it during the F15 atleast those that are considering as an default to catch any potential usability issues it did not eat any babes from me when I ran it. I think we should also ask ourselves these questions.. What benefit will this switch bring to the novice desktop end users? Will the novice desktop end user ever take advantages of any of the features that btrfs brings? Just my 0.0.2 cents.. JBG -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, Feb 22, 2011 at 5:07 PM, Lennart Poettering mzerq...@0pointer.de wrote: On Tue, 22.02.11 14:51, Josef Bacik (jo...@toxicpanda.com) wrote: Hello, So we're getting close to having a working fsck tool so I wanted to take the opportunity to talk about the future of BTRFS in Fedora. Coming up in F15 we're going to have the first release of Fedora where we don't need the special boot option to have the ability to format you filesystem as BTRFS. This is in hopes that we can open it up for wider testing before possibly making it the default filesystem. I realize we're in the early stages of F15, but since filesystems are big and important I'd like to get an idea of the amount of work that needs to still be done to get BTRFS in shape for being Fedora's default filesystem. So here are my goals 1) Fedora 16 ships with BTRFS as the default root filesystem. Hmm, what are your plans regarding hierarchy layout for this? My hope is that one day we can ship a read-only root dir by default, or more specifically a btrfs file system with three subvolumes in it: one read-only one mounted to /, and two writable ones mounted to /home and /var, with /tmp mounted from tmpfs. Yeah the hope is to separate out various things into different subvolumes so we can have things that can be independently snapshottable things. We came a long way with supporting read-only root dirs and it should mostly work now. In F15 for example /etc/mtab is a symlink, and even smaller stuff like /etc/nologin got moved to /var/run, to make write accesses to /etc unnecessary during normal operation. /etc/resolv.conf is the only thing often updated that's left, but with NM using dnsmasq it's static too. By using btrfs subvolumes doing this kind of seperation into writable and non-writable subtrees we don't have to think anymore about the sizes for those file systems at install time, since they all live in the same fs. If this is adopted package managers and system configuration UIs would need to invoke mount / -o rw,remount before doing their work. So, I'd like to see this implemented one day, maybe the adoption of btrfs is the right time to push this through too? I have not filed a feature page for this, as I am not sure I want to push this on F16, and I don't even know if people in general are onboard with this idea. The benefits of this are mostly security and robustness since we know that the actual subvolume the OS is booted from is always in a consistent state during operation and cannot normally be changed. And of course, we get all kind of magic by doing this because we can easily use snapshots to roll back system upgrades while leaving /home completely untouched. Sorry for hijacking your thread like this, It's cool, just hilights all the cool stuff we can do with BTRFS :). Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
2011/2/22 Jóhann B. johan...@gmail.com: On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: So what are your thoughts? Thanks, Will there be any performance penalties making this move? Who knows, thats what testing is for :). There are some things that suck with BTRFS, but so it goes with filesystems. I ran it on my workstation when F13 came out with updates+snapshots and while updating my desktop responsiveness was well not so good. Novice end users tools ( GUI ) and policy to clean those snapshots and perhaps storing those snapshots in it's own directory and or partition rather than directly on / need be in place before we implement this as an default. They would be helpful, but not necessary. I suggest people try it during the F15 atleast those that are considering as an default to catch any potential usability issues it did not eat any babes from me when I ran it. I think we should also ask ourselves these questions.. What benefit will this switch bring to the novice desktop end users? Various things, better data integrity to start with, and if you install the yum-fs-snapshot you have the ability to rollback easily. Will the novice desktop end user ever take advantages of any of the features that btrfs brings? At the very least you get better data integrity via our checksumming and duplicate metadata on single spindle fs's. Thanks, Josef -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, 2011-02-22 at 14:51 -0500, Josef Bacik wrote: 2) Fedora 16 ships without LVM as the volume manager and instead use BTRFS's built in volume management, again just for the default. In my personal opinion, this is a poor design decision. Yes, BTRFS can do a lot of volume-y things, and these are growing by the day, but I don't want my filesystem replacing a full volume manager and I am concerned that this will lead to less testing and exposure to full LVM use within the Fedora community. Instead, I'd like to counter-propose that everything stay exactly as it is, with users being able to elect to switch to BTRFS (sub)volumes if they are interested in doing so. Should the switch to BTRFS by default happen, this will be one more thing I will have to fix immediately during installation. The list grows longer and longer over time - please don't make this change. Jon. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Plans for BTRFS in Fedora
On Tue, Feb 22, 2011 at 14:51:50 -0500, Josef Bacik jo...@toxicpanda.com wrote: 3) All the various little tools that we have for putting together LiveCD's that are very ext* centered. I've not even looked at this yet, but I assume it's going to be kind of a pain. I like to see live CDs just use squashfs directly and not a squashed copy of an ext3/4 image. However this will break using dd to copy the image for installs which will slow things down. Further however, there is at least some interest in having live CDs and ananconda share the same install method, so this might be happening in any case. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel