> I see that ahc is storing its information in swift. That's clever, but if
> Ironic provided a blob store for each node, would that be better?

I can try to answer that. The initial implementation was doing that
but, AHC collect a fine-grained amount of data, e.g:

* it runs benchmark on all disks using different IO size. Read being
default but you can also test writing (disabled by default because
it's destructive)

* It tests memory performance with different blocks sizes (1K, 4K, 1M,
16M, 128M, 1G, 2G, ...)

* It stress test each physical CPU (and all at once)

So a machine with many disks, many CPUs and a lot of memory can
generate a really big amount of data. For that reason storing the blob
in Swift and just link the ID in Ironic is preferable.


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe

Reply via email to