Le Wed, 14 Nov 2018 22:26:51 +0100
Vojtech Pavlik <vojtech@xxxxxxxx> écrivait:
> >
> > So you're basically telling me that using md RAID 1 could possibly
> > be much slower than a Broadcom 94xx RAID controller for NVMe drives?
> > Interesting.
>
> It really depends on where the bottleneck is. But yes, it will be
> slower, because 'md' has no battery backed memory and thus has to
> limit the parallelism of the drives to achieve the same consistency
> guarantees.
OK. Benchmarking will tell, I suppose...
>
> > > So, make your own decision based on your usecase. :)
> >
> > Use case is : as fast as possible, but doesn't endanger the 600 TB
> > of data, because duh, 600 TB is quite a lot :)
>
> As much as I try to eschew hardware RAID, and push md for the sake of
> openness, in this case I believe it's the right solution for
> protecting against the NVMe SSD failure.
I'm waiting for my first batch of NVMe-capable RAID controllers to see
how they fare. I hope they don't limit throughput too much.
--
------------------------------------------------------------------------
Emmanuel Florac | Direction technique
| Intellique
| <eflorac@xxxxxxxxxxxxxx>
| +33 1 78 94 84 02
------------------------------------------------------------------------
Attachment:
pgpHPnlz5wY8c.pgp
Description: OpenPGP digital signature
