There’s more to it than bottlenecking.

RAS, man.  RAS.

> On Jul 12, 2024, at 3:58 PM, John Jasen <[email protected]> wrote:
> 
> How large of a ceph cluster are you planning on building, and what network 
> cards/speeds will you be using?
> 
> A lot of the talk about RAID HBA pass-through being sub-optimal probably 
> won't be your bottleneck unless you're aiming for a large cluster at 100Gb/s 
> speeds, in my opinion. 
> 
> On Fri, Jul 12, 2024 at 12:02 PM Drew Weaver <[email protected] 
> <mailto:[email protected]>> wrote:
>> Okay it seems like we don't really have a definitive answer on whether it's 
>> OK to use a RAID controller or not and in what capacity.
>> 
>> Passthrough meaning:
>> 
>> Are you saying that it's OK to use a raid controller where the disks are in 
>> non-RAID mode?
>> Are you saying that it's OK to use a raid controller where each disk is in 
>> its own RAID-0 volume?
>> 
>> I'm just trying to clarify a little bit. You can imagine that nobody wants 
>> to be that user that does this against the documentation's guidelines and 
>> then something goes terribly wrong.
>> 
>> Thanks again,
>> -Drew
>> 
>> 
>> -----Original Message-----
>> From: Anthony D'Atri <[email protected] <mailto:[email protected]>> 
>> Sent: Thursday, July 11, 2024 7:24 PM
>> To: Drew Weaver <[email protected] <mailto:[email protected]>>
>> Cc: John Jasen <[email protected] <mailto:[email protected]>>; 
>> [email protected] <mailto:[email protected]>
>> Subject: Re: [ceph-users] Repurposing some Dell R750s for Ceph
>> 
>> 
>> 
>> > 
>> > Isn’t the supported/recommended configuration to use an HBA if you have to 
>> > but never use a RAID controller?
>> 
>> That may be something I added to the docs.  My contempt for RAID HBAs knows 
>> no bounds ;)
>> 
>> Ceph doesn’t care.  Passthrough should work fine, I’ve done that for tends 
>> of thousands of OSDs, albeit on different LSI HBA SKUs.
>> 

_______________________________________________
ceph-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to