When the OSD is prepared, systemd-udev kicks in since the device has
been exclusively opened and thus reverts the permissions to root:disk.
We need a udev rule to force re-applying the correct ceph permission so
we can consume the disk.
Closes: https://github.com/rook/rook/issues/8942
Signed-off-by: Sébastien Han <seb@redhat.com>
The volume replication CRDs are an external component, not owned by Rook.
Therefore, they should be installed as any other independent component
in case the admin will install other consumers of the volumereplication CRDs
in the future in addition to Rook and the CSI driver.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The canary tests were still picking up the tag from operator.yaml
and toolbox.yaml instead of the new test local-build tag.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 6f48dce3f5)
This is a starting step to be able to generate common.yaml from Helm
charts. For right now, we merely want to be able to determine when the
rendered output of the Helm chart changes.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
Try to avoid the following:
```
error pulling image configuration: received unexpected HTTP status: 500 Internal Server Error
```
Signed-off-by: Sébastien Han <seb@redhat.com>
I've seen cases were the CI needs a few more seconds to delete and
object store. When logging in the runner, the object store is gone and
the timing matches too with the runner's logs (comparing with the
operator's logs).
Signed-off-by: Sébastien Han <seb@redhat.com>
We have seen cases where the signed certificate used for the RGW does not
contain the internal DNS endpoint, resulting in the health check to fail
since the certificate is not valid for this domain.
People consuming the gateways by external clients and for specific
domains do not necessarily have the internal DNS configured in the
certificate.
So let's be a bit more flexible and simply ensure a connectivity check
and bypass the certificate validation.
Also, this is fixing the tls code in newS3Agent and adds unit tests.
Closes: #8663
Signed-off-by: Sébastien Han <seb@redhat.com>
The `mgr host ls` looks like:
`"addr": "10.1.0.27/fv-az244-362", "hostname": "fv-az244-362"`
Previously, we were picking `addr` which contains both IP and hostname
where the k8snode has the hostname only. So let's use the Hostname from
the mgr command output instead.
Signed-off-by: Sébastien Han <seb@redhat.com>
In general, make the output from the object store test easier to follow
and debug.
- Make sure assertions are associated with the appropriate subtest.
- Remove `t.Helper()` from complex helpers.
- Add info to log messages.
- Move long parts of (sub-)test names to log messages.
- Move some reused functions back to ceph_base_object_test.go
- Refactor runObjectE2ETestLite() to be suitable for use as the second
store created in the "run a second object store" test.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
We have new jobs now:
* one that runs both smoke and object on the next Pacific version
* one that runs both smoke and object on Ceph master
* one that tests the upgrade from the current pacific stable to the
pacific devel
* one that tests the upgrade from the current octopus stable to the
octopus devel
Signed-off-by: Sébastien Han <seb@redhat.com>
In Rook v1.8 the min version of K8s supported is updated to 1.16.
Users running on older versions of K8s are recommended to update
to 1.16 or newer before updating to Rook v1.8.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The flex driver has been fully deprecated and thus removed from Rook.
Before upgrading to v1.8, users will need to convert existing flex volumes
from flex to csi volumes.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Simplify the test documentation on running the integration tests
to clarify that any K8s cluster can be used, and only support
minikube in the instructions.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Multiple CephObjectStores should be able to run at the same time. In CI,
test this by:
1. Create the primary object store under test (store A) (this was
already done)
2. Verify that store A starts and becomes ready (already done but is
moved here)
3. Create a second object store (store B) (new)
4. Verify that store B starts and becomes ready (new)
5. Delete store B, and ensure it is deleted (new)
6., etc. Continue to verify primary store A operations as before (this
was already done)
This is necessary to ensure that the Rook operator is using
`radosgw-admin` and the RGW Admin Ops API properly to support multiple
ObjectStores running at once.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
The integration tests must always be run against the local
build of rook, and an image should never be pulled from dockerhub.
To prevent pulling a release or master tag, the local build
will use a tag specific to the build and not ever published
elsewhere.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit a8a40428b0)
If the cluster where the rgw is started is secondary and not primary,
trying to create the admin ops user will fail with:
```
Please run the command on master zone.
Performing this operation on non-master zone
leads to inconsistent metadata between zones
```
So we need to force the creation regardless, it is fine the creation will
return UserAlreadyExist and then we just read the current user.
Closes: https://github.com/rook/rook/issues/8671
Signed-off-by: Sébastien Han <seb@redhat.com>
The MDS core team suggested with deploy the MDS daemon first and then do
the filesystem creation and configuration. Reversing the sequence lets
us avoid spurious FS_DOWN warnings when creating the filesystem.
Closes: #8745
Signed-off-by: Sébastien Han <seb@redhat.com>
If the region is specified in the storage class of OBC, use that in the
newS3agent() than using constant "us-east-1".
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
The `*SuiteMinimalTestVersion` vars in
`tests/integration/ceph_base_deploy_test.go` are no longer needed with
Jenkins no longer being used. Remove them.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
The CephSmokeSuite is becoming quite large and long, and most of the
length is now related to the object e2e tests. Separate the full object
e2e test into CephObjectSuite, and only test the object 'lite' test in
the CephSmokeSuite.
Resolves#8714
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
This commit is a large refactor on how the operator starts, stops and
how it starts various sub-components such as the ceph-csi driver. It
also refines the way we cancel orchestrations. We don't use breakpoints
anymore but send our self a SIGUP to reload our controller runtime
manager.
The reload will happen under different circonstances like:
* a new adminission controller secret is created/deleted/changed
* a CephCluster CR is edited
As mentioned earlier, the csi driver now has its own controller, just
like flex. It reacts to change in the operator config map for particular
ROOK_CSI_ fields.
A second new controller for the operator's general config has been
created, it manages:
* the logging level
* the ceph CLI command timeout
* the discovery daemon
The operator reacts much more rapidly to cancellation events by stopping
the manager's context and reloading it.
Signed-off-by: Sébastien Han <seb@redhat.com>
Let's force v16.2.5 since the CI is broken with 16.2.6. This gives us
time to continue to merge work and work on fixing deployments with
16.2.6 in parallel.
Signed-off-by: Sébastien Han <seb@redhat.com>
We should not use .items[0].metadata.name if the array length is 0. This
is the case when nothing has been initialized yet. Instead, we should
use .items[*].metadata.name, the wildcard ensures to always return 0
even if nothing is present yet.
Fixes: https://github.com/rook/rook/issues/8676
Signed-off-by: Sébastien Han <seb@redhat.com>
We just need to wait longer when the status is not ready. We needed
another sleep otherwise the status was never nil and the loop went too
fast. See:
```
2021-09-02 16:29:11.372249 I | integrationTest:
2021-09-02 16:29:11.374427 I | integrationTest:
2021-09-02 16:29:11.377764 I | integrationTest:
2021-09-02 16:29:11.379950 I | integrationTest:
2021-09-02 16:29:11.382084 I | integrationTest:
2021-09-02 16:29:11.385383 I | integrationTest:
2021-09-02 16:29:11.388499 I | integrationTest:
2021-09-02 16:29:11.391301 I | integrationTest:
2021-09-02 16:29:11.393545 I | integrationTest:
2021-09-02 16:29:11.396249 I | integrationTest:
```
Signed-off-by: Sébastien Han <seb@redhat.com>
Signed-off-by: Sébastien Han <seb@redhat.com>
When an OSD reaches OSD_NEARFULL state,
we have to manually increase the PVC volume claim
or manually increase the count of OSDs in the device set
Added a script auto-grow-storage.sh which will
i)automatically increase claim volume
ii)automatically add number of OSDs
Closes: https://github.com/rook/rook/issues/6101
Signed-off-by: parth-gr <paarora@redhat.com>
The Cassandra tests are failing in the CI, although they pass when
run locally on a developer cluster. Until the issue can be tracked
down we need to disable these checks to get to a green CI again.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Sébastien Han
Travis Nielsen
Blaine Gardner
Blaine Gardner
Satoru Takeuchi
Juan Miguel Olmo Martínez
Jiffin Tony Thottan
subhamkrai
parth-gr