[talk] S3 Work-a-likes

[Edward Capriolo]

On Thursday, November 3, 2016, Pete Wright <pete at nomadlogic.org> wrote:

>
>
> On 11/3/16 8:48 AM, Mark Saad wrote:
>
>>
>>
>>
>>
>>
>> On Thursday, November 3, 2016 10:58 AM, Siobhan Lynch <slynch2112 at me.com>
>>> wrote:
>>>
>>>>
>>>>
>>>
>>>
>>>  On Nov 3, 2016, at 10:38 AM, Mark Saad <mark.saad at ymail.com> wrote:
>>>>
>>>>  All
>>>>  Last night there was a discussion about s3 work-a-likes. Here are the
>>>> two
>>>>
>>> that I know of.
>>>
>>>>
>>>>  1. LeoFS .
>>>>     Written in Erlang, and runs well enough on FreeBSD, Solaris,
>>>> Illumos ,
>>>>
>>> Linux and maybe even  Windows.
>>>
>>>>
>>>>  Written by rakuten and used in production on a number of things.
>>>> Including
>>>>
>>> Project-Fifo.net, Rakuten.com
>>>
>>>>  and a number of other systems.
>>>>
>>>>  http://leo-project.net/leofs/
>>>>
>>>>  2. Minio .
>>>>    Written Go , and claims to be supported on many platforms .
>>>>  Its fairly new but aims to be fully compatable.
>>>>  https://minio.io
>>>>
>>>>
>>>>  Also fun things with LeoFS, it can run inside of a Zone or a Jail.
>>>>
>>>>
>>>>
>>>
>>> Ceph has radosgw which is a REST gateway to emulate s3.
>>>
>>
>>
>>>
>>
>> Trish
>>   I want to know why everyone is talking about ceph ? Not that I am in
>> love with LeoFS .
>> But ceph has lots of issues on not linux.
>>
>>
> tl;dr version: it's shite.  the longer story...it's shite and focused on
> linux pretty much.
>
> i kid i kid - it's def some interesting technology, but even before RedHat
> acquired Inktank there was a ton of linux specific foo baked into the code.
>
> and even if you are running it on linux it's really unstable IMHO outside
> of some specific use-cases.  but even then, using it as a block storage
> backend for openstack for example, it is definitely not for the faint of
> heart to implement and maintain at scale.
>
> just my two bits - i'm sure there are plenty of counterpoints out there -
> but i've often put lots of hope that ceph will mature into a product as it
> would have made my life *much* easier only to be disappointed.
>
> -p
>
>
> --
> Pete Wright
> pete at nomadlogic.org
> nomadlogicLA
>
> _______________________________________________
> talk mailing list
> talk at lists.nycbug.org
> http://lists.nycbug.org/mailman/listinfo/talk
>

Also s3 is not a mountable filesytem. It is not a zfs, lusterfs, or
glusterfs not looking to implement posix locking etc.

Amazons product prospective is key. Instead of attemping to solve the
distributed mountable lockable filesystem problem, they understood that
people were making large piles of mostly large file immutable data.

Compliance and regulation want to keep all the data, bi teams wanting to do
year over year reporting want to keep the data. The built a solution not
attempted to solve the holy grail of big distributed file system with
posix.

Put/get big scale. Once they lock that piece they add features. Auto
retention via meta data, put get generate events to kinesis, reduced
storage lower cost, glacier only pay to take out.


-- 
Sorry this was sent from mobile. Will do less grammar and spell check than
usual.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.nycbug.org:8443/pipermail/talk/attachments/20161103/958d2d58/attachment.htm>