Since each entry in a L2 table represents a cluster block within the
QCOW image, we can compute deltas by comparing L2 tables. This fixes
an issue where snapshots were not being properly thin-provisioned
during the live-migration execute process.
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
This will be used in the next commit where the list-snaps state
machine is tweaked to compate L2 tables between snapshot revisions
to determine deltas.
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
This will be used by a refactored list-snaps state machine to compute
the L2 table cluster for image offsets. The L1 and L2 lookup tables
can now be represented in a single structure.
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
mgr/dashboard: Fix for incorrect validation in rgw user form
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>
Reviewed-by: Volker Theile <vtheile@suse.com>
`DaemonDescription.status` is annotated
as `Optional[int]` and thus can be `None`
This is a conflict between c95ba878c6
and 01f60cf4e0
Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
cephadm: latest stable release is now pacific
Reviewed-by: Juan Miguel Olmo Martínez <jolmomar@redhat.com>
Reviewed-by: Michael Fritch <mfritch@suse.com>
* fetch_config() before mkfs and starting osd
for populating settings related to booting and transport layer
before it starts.
* set fsid read from monitor before mkfs
it's crucial to mkfs if osd is supposed to retrieve the fsid
from monitor.
Signed-off-by: Kefu Chai <kchai@redhat.com>
just for waiting for monmap and config from mon. crimson-osd needs this
for populating settings related to booting and transport layer before it
starts.
Signed-off-by: Kefu Chai <kchai@redhat.com>
so, for instance, if we want to connect to monitor without bind to any
address, we can try to use the v2 addresses advertised in monmap or
local settings, instead of being unable to connect to mon because we are
using an `entity_addr_t::TYPE_NONE` address which is returned by
`entity_addrvec_t::front()` if the addrvec is empty.
see also AsyncMessenger::should_use_msgr2().
Signed-off-by: Kefu Chai <kchai@redhat.com>
The rgw users create form doesnt validate the username correctly if the username is a tenated one. For eg. Consider there is a user called tenate$sample. Now I am trying to create another user and I entered tenate$sample as username. But it doesn't async validate the username as existing. Instead it just shows the green tick and once the submit button is clicked it'll show the user as existing.
Fixes: https://tracker.ceph.com/issues/48907
Signed-off-by: Nizamudeen A <nia@redhat.com>
The expiration timestamp was hard-coded as
2020-12-31T19:14:44.180394
which is now in the past. Instead, use a timestamp
90 minutes in the future.
Fixes: https://tracker.ceph.com/issues/48919
Signed-off-by: Marcus Watts <mwatts@redhat.com>
The existing logic uses "python -m zipfile" to unzip files.
This will (most likely) fail on CentOS 8-Stream , where python
defaults to 'unset' (see man unversioned-python).
So: try harder: try unzip, python3, and python in that order,
to find something that can unzip files.
Fixes: https://tracker.ceph.com/issues/48921
Signed-off-by: Marcus Watts <mwatts@redhat.com>
Fixes a bug introduced earlier in the patchset. It's just too hard to go
back and fix it up with the right commit.
Signed-off-by: Joao Eduardo Luis <joao@suse.com>
