1 |
Am Thu, 08 May 2014 19:57:34 +0800 |
2 |
schrieb William Kenworthy <billk@×××××××××.au>: |
3 |
|
4 |
> On 05/07/14 07:51, Marc Joliet wrote: |
5 |
> > Am Wed, 07 May 2014 06:56:12 +0800 |
6 |
> > schrieb William Kenworthy <billk@×××××××××.au>: |
7 |
> > |
8 |
> >> On 05/06/14 18:18, Marc Joliet wrote: |
9 |
> >>> Hi all, |
10 |
> >>> |
11 |
> >>> I've become increasingly motivated to convert to btrfs. From what I've seen, |
12 |
> >>> it has become increasingly stable; enough so that it is apparently supposed to |
13 |
> >>> become the default FS on OpenSuse in 13.2. |
14 |
> >>> |
15 |
> >>> I am motivated by various reasons: |
16 |
> >> .... |
17 |
> >> |
18 |
> >> My btrfs experience: |
19 |
> >> |
20 |
> >> I have been using btrfs seriously (vs testing) for a while now with |
21 |
> >> mixed results but the latest kernel/tools seem to be holding up quite well. |
22 |
> >> |
23 |
> >> ~ 2yrs on a Apple/gentoo laptop (I handed it back to work a few months |
24 |
> >> back) - never a problem! (mounted with discard/trim) |
25 |
> > That's one HDD, right? From what I've read, that's the most tested and stable |
26 |
> > use case for btrfs, so it doesn't surprise me that much that it worked so well. |
27 |
> > |
28 |
> Yes, light duty using the builtin ssd chips on the motherboard. |
29 |
|
30 |
SSD chips on the motherboard? I just did a quick googled (well, "duckduckwent" |
31 |
would be more accurate; funny how language works), do you mean something like |
32 |
the Intel Z68 chipset? |
33 |
|
34 |
> >> btrfs on a 128MB intel ssd (linux root drive) had to secure reset a few |
35 |
> >> times as btrfs said the filesystem was full, but there was 60G+ free - |
36 |
> >> happens after multiple crashes and it seemed the btrfs metadata and the |
37 |
> >> ssd disagreed on what was actually in use - reset drive and restore from |
38 |
> >> backups :( Now running ext4 on that drive with no problems - will move |
39 |
> >> back to btrfs at some point. |
40 |
> > All the more reason to stick with EXT4 on the SSD for now. |
41 |
> I have had had very poor luck with ext anything and would hesitate it to |
42 |
> recommend it except for this very specific case where there is little |
43 |
> alternative - reiserfs is far better on platters for instance. |
44 |
|
45 |
I, like Stefan, am interested in precisely what kind of negative experiences |
46 |
you've had with ext*. |
47 |
|
48 |
I used to use reiserfs (from waaay back when I still used SuSE), but the only |
49 |
remnant of that is a broken external HDD that I want to attempt a ddrescue on |
50 |
someday (really the only reason I still keep around reiserfs support in my |
51 |
kernel). The only thing I really miss is tail packing of small files; my actual |
52 |
disk usage grew noticeably after my switch to ext4. But ext* have the distinct |
53 |
advantage of being used pretty much everywhere, which, as we all know, is an |
54 |
important factor in finding bugs. Reiserfs, in comparison, is AFAIK |
55 |
unmaintained now (of course, it's maintained in the sense that the existing |
56 |
code is kept working, but that's beside the point). |
57 |
|
58 |
> > [snip interesting but irrelevant ceph scenario] |
59 |
> Its relevant because it keeps revealing bugs in btrfs by stressing it - |
60 |
> one of those reported by me to ceph was reported upstream by the ceph |
61 |
> team and fixed last year - bugs still exist in btrfs ! |
62 |
|
63 |
Sorry, I read "ceph" and immediately thought "OK, clustering file system, way |
64 |
outside of experience" and didn't realise you were talking about outright bugs |
65 |
in btrfs (I kind of assumed a situation similar to btrfs and swap files: one |
66 |
piece of software (e.g., the kernel swap code) making assumptions that don't |
67 |
hold for btrfs). |
68 |
|
69 |
> >> 3 x raid 0+1 (btrfs raid 1 with 3 drives) - working well for about a month |
70 |
> > That last one is particularly good to know. I expect RAID 0, 1 and 10 to work |
71 |
> > fairly well, since those are the oldest supported RAID levels. |
72 |
> > |
73 |
> >> ~10+ gentoo VM's, one ubuntu and 3 x Win VM's with kvm/qemu storage on |
74 |
> >> btrfs - regular scrubs show an occasional VM problem after system crash |
75 |
> >> (VM server), otherwise problem free since moving to pure btrfs from |
76 |
> >> ceph. Gentoo VM's were btrfs in raw qemu containers and are now |
77 |
> >> converted to qcow2 - no problems since moving from ceph. Fragmentation |
78 |
> >> on VM's is a problem but "cp --reflink vm1 vm2" for vm's is really |
79 |
> >> really cool! |
80 |
> > That matches the scenario from the ars technica article; the author is a huge |
81 |
> > fan of file cloning in btrfs :) . |
82 |
> > |
83 |
> > And yeah, too bad autodefrag is not yet stable. |
84 |
> Not that its not stable but that it cant deal with large files that |
85 |
> change randomly on a continual basis like VM virtual disks. |
86 |
|
87 |
Oh, I thought it was still considered new and "unpolished" (I did not mean |
88 |
buggy!). |
89 |
|
90 |
> >> I have a clear impression that btrfs has been incrementally improving |
91 |
> >> and the current kernel and recovery tools are quite good but its still |
92 |
> >> possible to end up with an unrecoverable partition (in the sense that |
93 |
> >> you might be able to get to some of the the data using recovery tools, |
94 |
> >> but the btrfs mount itself is toast) |
95 |
> >> |
96 |
> >> Backups using dirvish - was getting an occasional corruption (mainly |
97 |
> >> checksum) that seemed to coincide with network problems during a backup |
98 |
> >> sequence - have not seen it for a couple of months now. Only lost whole |
99 |
> >> partition once :( Dirvish really hammers a file system and ext4 usually |
100 |
> >> dies very quickly so even now btrfs is far better here. |
101 |
> > I use rsnapshot here with an external hard drive formatted to EXT4. I'm not |
102 |
> > *that* worried about the FS dying, more that it dies at an inopportune moment |
103 |
> > where I can't immediately restore it. |
104 |
> > |
105 |
> > [again, snip interesting but irrelevant ceph scenario] |
106 |
> as I said above - if it fails under ceph, its likely going to fail under |
107 |
> similar stresses using other software - I am not talking ceph bugs (of |
108 |
> which there are many) but actual btrfs corruption. |
109 |
|
110 |
Got it, thanks! |
111 |
|
112 |
> >> I am slowly moving my systems from reiserfs to btrfs as my confidence in |
113 |
> >> it and its tools builds. I really dislike ext4 and its ability to lose |
114 |
> >> valuable data (though that has improved dramaticaly) but it still seems |
115 |
> >> better than btrfs on solid state and hard use - but after getting burnt |
116 |
> >> I am avoiding that scenario so need to retest. |
117 |
> > Rising confidence: good to hear :) . |
118 |
> > |
119 |
> > Perhaps this will turn out similarly to when I was using the xf86-video-ati |
120 |
> > release candidates and bleeding edge gentoo-sources/mesa/libdrm/etc. (for 3D |
121 |
> > support in the r600 driver): I start using it shortly before it starts truly |
122 |
> > stabilising :) . |
123 |
> > |
124 |
> More exposure, more bugs will surface and be fixed - its getting there. |
125 |
|
126 |
I sure hope so, I'll be making the switch either tomorrow or Monday :-) . |
127 |
|
128 |
> BillK |
129 |
|
130 |
-- |
131 |
Marc Joliet |
132 |
-- |
133 |
"People who think they know everything really annoy those of us who know we |
134 |
don't" - Bjarne Stroustrup |