[lustre-discuss] mkfs.lustre for ZFS draid

[Thomas Roth]

Yeah, maybe I did not understand draid in all aspects...
The reason for 7 OSTs is simply that I want to replace old OSS, not add new ones. And the existing servers are setup like this, with 7 OSTs each.
In future, we will probably be forced to go the other way, making smaller OSTs (aka OSTs with less disks), to keep the number of OST indices per NID, 
but that is certainly not the use case for draid.

Regards,
Thomas

On 2/6/25 15:40, Laura Hild wrote:
>> But how can I address them when running mkfs.lustre, if I still want to get 7 OSTs out of it?
> 
> When we started using draid, we went from having six ten-disk OSTs per OSS to one sixty-disk OST per OSS.  I'm unsure why one would want seven OSTs on a single pool, being that if the pool becomes degraded all the OSTs would with it.  It would be as if you made a pool with seven raidz2 vdevs and created seven datasets.  The datasets do not correspond to individual vdevs, but rather all of the datasets use all of the vdevs.
>