hi casey,
well, I *think* that our current redis logic in Samarah's filter is/was
storing object xattrs only to simplify getting something working. I don't
think that behavior is based on a the original d4n, or would be correct
independent of cache backend
but as you state, if the cache backend is redis, it should store the
xattrs, &c I don't have an intuition about redis notation, but that seems
sensible to me
Matt
On Tue, Aug 8, 2023 at 1:20 PM Casey Bodley <cbodley(a)redhat.com> wrote:
thanks Matt,
the d4n filter is currently storing all of the object's xattrs in the
redis directory, so i wasn't sure we needed them in the cache backend
too. it sounds like we don't actually want the directory to store
those
we're trying to decide what representation to use for the redis cache
backend. if we only store data in a redis key, we can use
SETRANGE/GETRANGE to operate on ranges. if we include object
attributes in the same key, i think we have to use key/value
interfaces like HSET/HGET instead. so if the cache needs to do both,
we probably want each object to use a different redis key for its data
and metadata
On Tue, Aug 8, 2023 at 12:44 PM Matt Benjamin <mbenjami(a)redhat.com> wrote:
I don't think it's ever been the case that the d4n directory (redis)
uniquely stores all of what RGW considers attributes. Just "it's own
attributes" (attributes relevant to cache operation and strategy; e.g.,
data location, hit rates, anything else that eventually becomse part of the
cache workflow), so far as I know. I think other attributes such as user
metadata should be stored on the data cache.
Matt
On Tue, Aug 8, 2023 at 11:54 AM Samarah Uriarte <suriarte(a)redhat.com>
wrote:
>
> Hello everyone,
>
> Casey had raised a question recently regarding the design of the Cache
Driver.
He was under the impression that the D4N directory should be
storing object metadata (and attributes) while the Cache Driver should be
storing data only.
>
> Currently, the D4N Directory only stores metadata specific to
directory-related operations, which is why S3 attributes like bucket size,
mtime, and others are not added to the object's directory entry. Dan has
also mentioned the prospect of moving SAL attributes into the directory
rather than the cache backend; so this is worth discussing in more detail.
>
> If you have any thoughts or additional comments on this topic, please
let me
know.
Sincerely,
Samarah Uriarte
--
Matt Benjamin
Red Hat, Inc.
315 West Huron Street, Suite 140A
Ann Arbor, Michigan 48103
http://www.redhat.com/en/technologies/storage
tel. 734-821-5101
fax. 734-769-8938
cel. 734-216-5309
--
Matt Benjamin
Red Hat, Inc.
315 West Huron Street, Suite 140A
Ann Arbor, Michigan 48103
tel. 734-821-5101
fax. 734-769-8938
cel. 734-216-5309