1 |
The HTML...it hurts my eyes... :) |
2 |
|
3 |
On Thu, Sep 18, 2014 at 4:24 AM, J. Roeleveld <joost@××××××××.org> wrote: |
4 |
> |
5 |
> On Wednesday, September 17, 2014 08:56:28 PM James wrote: |
6 |
> |
7 |
>> Alec Ten Harmsel <alec <at> alectenharmsel.com> writes: |
8 |
> |
9 |
>> > As far as HDFS goes, I would only set that up if you will use it for |
10 |
>> > Hadoop or related tools. It's highly specific, and the performance is |
11 |
>> > not good unless you're doing a massively parallel read (what it was |
12 |
>> > designed for). I can elaborate why if anyone is actually interested. |
13 |
> |
14 |
|
15 |
FYI - one very big limitation of hdfs is its minimum filesize is |
16 |
something huge like 1MB or something like that. Hadoop was designed |
17 |
to take a REALLY big input file and chunk it up. If you use hdfs to |
18 |
store something like /usr/portage it will turn into the sort of |
19 |
monstrosity that you'd actually need a cluster to store. |
20 |
|
21 |
> |
22 |
> My current understanding is: |
23 |
> |
24 |
> - ZFS is production ready, but due to licensing issues, not included in the |
25 |
> kernel |
26 |
> |
27 |
> - BTRFS is included, but not yet production ready with all planned features |
28 |
> |
29 |
|
30 |
Your understanding of their maturity is fairly accurate. They also |
31 |
aren't 100% moving in the same direction - btrfs aims more to be a |
32 |
general-purpose filesystem replacement especially for smaller systems, |
33 |
and zfs is more focused on the enterprise, so it lacks features like |
34 |
raid reshaping (who needs to add 1 disk to a raid5 when you can just |
35 |
add 5 more disks to your 30 disk storage system). |
36 |
|
37 |
I think btrfs has a bit more hope of being an ext4 replacement some |
38 |
day for both this reason and the licensing issue. That in no way |
39 |
detracts from the usefulness of zfs, especially for larger deployments |
40 |
where the few areas where btrfs is more flexible would probably be |
41 |
looked at as gimmicks (kind of like being able to build your whole OS |
42 |
from source :) ). |
43 |
|
44 |
> For me, Raid6-like functionality is an absolute requirement and latest I |
45 |
> know is that that isn't implemented in BTRFS yet. Does anyone know when that |
46 |
> will be implemented and reliable? Eg. what time-frame are we talking about? |
47 |
> |
48 |
|
49 |
I suspect we're talking months before it is really implemented, and |
50 |
much longer before it is reliable. Right now btrfs can write raid6, |
51 |
but it can't really read it. That is, it operates just fine until you |
52 |
actually lose a disk containing something other than parity, and then |
53 |
it loses access to the data. This code is only in the kernel for |
54 |
development purposes and nobody advocates using it for production. |
55 |
Most of the code in btrfs which is reliable has been around for years, |
56 |
like raid1 support, and obviously it will be years until the raid5/6 |
57 |
code reaches that point. I am using btrfs mainly because once that |
58 |
day comes it will be much easier to migrate to it from btrfs raid1 |
59 |
than from zfs (which has no mechanism for migrating raid levels |
60 |
in-place (that is, within an existing vdev) - you would need to add |
61 |
new drives to the pool, migrate the data, and remove the old drives |
62 |
from the pool, which is nice if you have a big stack of drives and |
63 |
spare sata ports lying around like you would in a SAN). |
64 |
|
65 |
-- |
66 |
Rich |