* The flush/evict won't work if target_max_bytes/objects are not configured.
* All client requests will be blocked only when target_max_bytes/objects
are reached. Hitting on cache_target_full_ratio will not block client
requests.
Signed-off-by: Yuan Zhou <yuan.zhou@intel.com>
Elaborates the note section a bit more as to why stopping I/O before
taking a snapshot is recommended and how to do it.
Signed-off-by: Nilamdyuti Goswami <ngoswami@redhat.com>
Object size can be specified when creating an image with the --order option,
as a number of bits in the size.
This patch is adding new option --object-size. This new option will specify
object size directly for example --object-size 2M.
It would be easier to use. --order is still present for backwards compatibility.
For simplicity, we are rounding up the object size to the nearest power of 2.
Fixes#12112
Signed-off-by: Vikhyat Umrao <vumrao@redhat.com>
rgw multitenancy
Buckets and users can now have a tenant associated with them. Buckets can be specified as [<tenant>:]<bucket> and users can be specified as [<tenant>$]<uid>. The existing users and buckets are part of the default global tenant.
This is first phase. Future work will add tighter integration (e.g., swift tenant mapping, s3 subdomain tenant mapping, etc.)
Reviewed-by: Pete Zaitcev <zaitcev@redhat.com>
Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
Replace existing index.rst with the beginnings of a comprehensive
developer guide.
Move the old index.rst to index-old.rst.
Signed-off-by: Nathan Cutler <ncutler@suse.com>
There is a typo in document "Monitoring OSDs and PGs" as below
An OSD was ``down``, was restared, and is now ``recovering``.
This patch corrects 'restared' to 'restarted'. The whole code base
is also searched to ensure there is no more 'restared'.
Signed-off-by: Yilong Zhao <accelazh@gmail.com>
ceph osd pool set $POOL scrub_min_interval N
ceph osd pool set $POOL scrub_max_interval N
ceph osd pool set $POOL deep_scrub_interval N
If N > 0, this value is used for the pool instead of
the corresponding global parameter from the config
(osd_scrub_min_interval, osd_scrub_max_interval or
osd_deep_scrub_interval).
Fixes: #13077
Signed-off-by: Mykola Golub <mgolub@mirantis.com>
Previous commits introduced the tenant infrastructure such as
rgw_user, but did not allow anyone to access it productively.
Done:
- radosgw-admin
- bucket creation, listing, deletion with non-empty tenant
- COPY
- ACLs
- Using colon for S3 with URL addressing
TODO:
- Fix S3 remapping of DNS so that period turns into colon in buckets,
possibly using typed endpoint domains, some assuming tenants
- Have Swift authentication to set tenant into URL, then pick it there
- Resolve leftover XXX
Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
