> Muwhahahahaha, that was a much needed laugh after the adrenaline rush of
> phasing in a cache tier in the wee hours of last night. ^o^
${subject} is my bad -;
A pretty much lol. Sorry for my typo!
> 1. It quite amuses me that a Redhat employee (or so one deduces from your
> mail address) goes to the community to ask about Ceph when the answers are
> available in-house so to speak (and seemingly quite some behind your
> pay-wall) as well as a google search away.
> http://tracker.ceph.com/issues/1576
BTW my all questions sent to this ML is *totally* unrelated to Red Hat.
Probably I should use different email address from the next.
I'm just interested in distributed filesystem, since I was the Lustre, the GPFS
guy for a long period and now the Ceph.
> 2. Other means to dedup with Ceph have also been considered, to not much
> avail:
> http://lists.ceph.com/pipermail/ceph-users-ceph.com/2015-April/000243.html
I read [1] and [2] and do understood Greg's comment [3].
If I've missed nothing, there is no discussion about deduplication feature (NOT
dedumplication or dumpling) except [1] and there is no that feature in the Ceph
and no consideration about that feature.
[1] http://tracker.ceph.com/issues/1576
[2]
http://docs.ceph.com/docs/hammer/rados/configuration/filesystem-recommendations/
[3] http://tracker.ceph.com/issues/1576#note-5
Some enterprise people want this feature [4].
There should be any kind of consideration on the sly, I guess.
[4] http://www.mellanox.com/blog/2015/06/ceph-is-a-hot-storage-solution-but-why/
> Lastly while deduplication would be very nice, there are least a dozen
> other features/bug fixes of indefinite higher priority that would need to
> go into Ceph to make it a more "enterprise level" product.
Yeah, you're right.
> On the top of my list is of course the lack of object checksums and
> resulting inability of Ceph to recover automatically and safely from scrub
> errors. All that redundancy, only for having to play Russian roulette when
> comes to fix data corruption.
Finally nobody wants *Russian roulette* -;
Cheers,
Shinobu
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
