Commit Graph
1370 Commits
Author SHA1 Message Date
Travis Nielsen 215758bfea Merge pull request #5795 from cybozu-go/ceph-fix-the-wrong-description-of-mon-placement
ceph: remove an outdated document of mon placement
2020-07-14 07:30:43 -06:00
Ahmad Nurus S 1659687d37 nfs: fix nfs issue on openshift
Signed-off-by: Ahmad Nurus S <prksu.sh@gmail.com>
2020-07-14 18:11:39 +07:00
Satoru Takeuchi 63d077ffde ceph: remove an outdated document of mon placement
Although the document describes we can't set Pod{,Anti}Affinity
to mon's placement, the current implementation permits to do so.

Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
2020-07-14 07:42:16 +00:00
Sébastien Han ccb52b84e6 ceph: configurable status checks and livenessprobe
This commit allows us to configure status check for each daemon:

* "mon": health check on the ceph monitors (quorum)
* "osd": health check on the ceph osds
* "status": ceph health status check

Each check is controlled by the following settings:

* disabled: whether to disable the check (default: false)
* internal: interval to run the check
* timeout: only valid for mons, is the timeout for unresponsive mon
before failling over.

Example to disable the status health check:

```yaml
healthCheck:
  daemonHealth:
    status:
      disabled: true
```

As part of that, pod's livenessprobe can now be configured via the
following settings:

* disabled: whether to enable or not
* probe: override the current probe in place by a new one

```yaml
healthCheck:
  livenessProbe:
    mon:
      disabled: true
```

Closes: https://github.com/rook/rook/issues/5772
Signed-off-by: Sébastien Han <seb@redhat.com>
2020-07-10 19:21:59 +02:00
Blaine Gardner 3678bd791a docs: ceph: add no support for changing networking
Clarify in the docs that changing networking modes after a Cpeh cluster
has been deployed is not supported and can result in a nonfunctional
cluster.

Signed-off-by: Blaine Gardner <blaine.gardner@suse.com>
2020-07-09 12:17:27 -06:00
Travis Nielsen ba590396e8 docs: remove references to xfs for ceph volumes
When mounting a ceph volume, xfs may cause issues with potential
deadlocks if the volume is on the same node as the OSDs. To avoid
this issue the default recommendation is ext4.

Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
2020-07-08 22:31:11 -06:00
Travis Nielsen dbae36b2bb ceph: update to the latest octopus release
Update the base operator image and cluster examples to use the latest
octopus release v15.2.4. Also cleanup some old examples and unit tests
that were still based on mimic.

Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
2020-07-06 12:11:47 -06:00
Sébastien Han 3f97e1561e Merge pull request #5748 from stuartpb/patch-1
ceph: document the `deviceClass` setting for OSDs
2020-07-06 11:20:45 +02:00
Satoru Takeuchi 48a87d192a ceph: fix a wrong link
`ceph-` suffix is necessary.

Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
2020-07-06 00:34:23 +00:00
Ahmad Nurus S d324f15ec4 nfs: add kubebuilder marker and fix some documentation
Signed-off-by: Ahmad Nurus S <prksu.sh@gmail.com>
2020-07-03 14:33:36 +07:00
Ahmad Nurus S 90da5ec7b9 nfs: add validation admission webhook using controller-runtime
Signed-off-by: Ahmad Nurus S <prksu.sh@gmail.com>
2020-07-03 14:33:36 +07:00
Ahmad Nurus S f7f8270f58 nfs: rewrite nfs operator controller to use controller-runtime
Signed-off-by: Ahmad Nurus S <prksu.sh@gmail.com>
2020-07-03 14:33:25 +07:00
Stuart P. Bentley 35f202bd95 ceph: document the deviceClass setting for OSDs
This functionality was merged in with #3696, but not documented.

Signed-off-by: Stuart P. Bentley <s@stuartpb.com>
2020-07-03 00:37:33 +00:00
Sébastien Han e4eaa91ede ceph: add rgw endpoint healthcheck
We have introduced a new goroutine to check the state of the rgw
endpoint. It will run every minute and perform operations on a bucket.
The success or failure will be reported as part of the status field of
the CephObjectStore CR.

A good status will look like:

status:
  endpointStatus:
    lastChanged: "2020-06-25T13:47:45Z"
    lastChecked: "2020-06-25T13:48:46Z"
  phase: Connected

A failed status:

status:
  endpointStatus:
    details: |-
      error creating bucket "rook-ceph-internal-s3-bucket-checker": RequestError: send request failed
      caused by: Put http://rook-ceph-rgw-my-store.rook-ceph:8080/rook-ceph-internal-s3-bucket-checker: dial tcp 10.108.189.148:8080: connect: connection refused
    health: ERROR

This check works for both converged and external modes. Note that the
CephObjectStore CRD has a new field called "externalRgwEndpoints" which
allows you to define a list of IP addresses pointing to rgws.

Closes: https://github.com/rook/rook/issues/5692
Signed-off-by: Sébastien Han <seb@redhat.com>
2020-07-02 16:34:10 +02:00
Sébastien Han 4636da67d3 Merge pull request #5679 from thotz/updateobcdoc
docs: update ROOK_OBC_WATCH_OPERATOR_NAMESPACE in obc doc
2020-07-01 11:15:50 +02:00
Madhu Rajanna cc97b94953 ceph: add upgrade documentation for csi
as we are moving from alpha to beta snapshot,
we need to cleanup alpha resources to have successful
upgrade

Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
2020-07-01 07:20:04 +05:30
Madhu Rajanna 3f46b437c6 ceph: remove CSI_ENABLE_SNAPSHOTTER from CSI config
removed CSI_ENABLE_SNAPSHOTTER from the configuration
to enable or disable csi-snapshotter as snapshot
is moved to beta and enabled by default in kubernetes
1.17+

Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
2020-07-01 07:20:04 +05:30
Madhu Rajanna 6699dba25a ceph: update sidecar images version
updated snapshotter from v1.x to v2.x
as its supports snapshot beta, the required
provisioner version to support snapshot beta
is v1.5.x+.

Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
2020-07-01 07:20:04 +05:30
Madhu Rajanna 2e973d03ec ceph: update snapshot doc as per beta version
Updated the snapshot usage doc as per new
snapshot beta version.

Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
2020-07-01 07:20:04 +05:30
Jiffin Tony Thottan 3bed313fbd docs: update ROOK_OBC_WATCH_OPERATOR_NAMESPACE in obc doc
Closes: 5675
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
2020-06-30 23:04:26 +05:30
Jiffin Tony Thottan a8d06ffd4f docs: mention about internal port consumed by rgw daemon
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
2020-06-25 23:59:03 +05:30
Travis Nielsen 6c0d4057b3 Merge pull request #5045 from vbnrh/rook-admission-controller
ceph: Adding support for admission controllers
2020-06-25 09:56:39 -06:00
Vineet Badrinath ce1003aef8 ceph: adds scripts and components to support admission controllers
adds deploy.sh script to deploy validatingwebhookconfiguration and create secrets.
adds new command ceph admission-controller to start webhook servers.
adds validation for various rook custom resources

Signed-off-by: Vineet Badrinath <vbadrina@redhat.com>
2020-06-24 14:59:00 +05:30
Sébastien Han 41dff79513 Merge pull request #5691 from cybozu-go/ceph-fix-the-wrong-document-of-tuneDeviceClass
ceph: fix the wrong document of tuneDeviceClass
2020-06-24 10:23:27 +02:00
Satoru Takeuchi c8c05ed7b9 ceph: fix the wrong document of tuneDeviceClass
The tuning parameter of slow device is defined as "tuneDeviceClass".
However, it's written in the document as "tuneSlowDeviceClass".

Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
2020-06-24 00:31:14 +00:00
Satoru Takeuchi 8ffe096616 ceph: fix a wrong link
There is a wrong link to the monitoring document.

Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
2020-06-23 22:31:35 +00:00
Sébastien Han c7f255a0e8 ceph: add the ability to set any pool property
We can now explicitly set any property on a given pool by using the new
Property field in the CephBlockPool Spec.

Also, this fixes the case where both `CephBlockPool` and `CephCluster`
are created at the same time. When Rook creates the pool, the cluster is
still being bootstrapped and the global option
`osd_pool_default_pg_autoscale_mode` has not bee set yet. So the pool
gets created but its `pg_autoscale_mode` property is set to `warn`
instead of `on`.

Closes: https://github.com/rook/rook/issues/5608V
Signed-off-by: Sébastien Han <seb@redhat.com>
2020-06-17 10:08:00 +02:00
Jared Watts c0ed19aa4b Merge pull request #5567 from prksu/design-nfs-controller-runtime
design: nfs operator using controller-runtime
2020-06-15 12:05:49 -07:00
Sébastien Han 3d8fd2ca5a Merge pull request #5630 from cybozu-go/add-osdpodspec-schedulername
ceph: add schedulerName property to storageClassDeviceSet
2020-06-15 10:25:57 +02:00
UMEZAWA Takeshi 6085ff3223 ceph: add schedulerName property to storageClassDeviceSet
Support schdulerName to use another scheduler for OSD pod placement.

Closes: https://github.com/rook/rook/issues/5536
Signed-off-by: UMEZAWA Takeshi <takeshi-umezawa@cybozu.co.jp>
Signed-off-by: Masayuki Ishii <masa213f@gmail.com>
2020-06-12 11:53:28 +09:00
Travis Nielsen 899e8e547e docs: clean ssds with blkdiscard
blkdiscard is the better tool for cleaning SSDs instead
of dd.

Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
2020-06-11 15:45:10 -06:00
Travis Nielsen 955dcb930e Merge pull request #5377 from alimaredia/wip-object-store-multisite
Foundation for Object Store Multisite CRs
2020-06-10 16:07:53 -06:00
Sébastien Han db17533435 docs: add design doc template
Rook already has a design process in place when introducing new
functionality. The only missing thing is a clear template to bring
consistency in our design documents.

This is inspired and shrunk version of
https://github.com/openshift/enhancements/tree/master/guidelines to add
some guidelines when it comes to writing a design document.

Signed-off-by: Sébastien Han <seb@redhat.com>
2020-06-10 19:05:14 +02:00
Ahmad Nurus S f5f6484869 docs: design nfs operator using controller-runtime
This commit is adding a design to implement controller-runtime in nfs operator

Signed-off-by: Ahmad Nurus S <prksu.sh@gmail.com>
2020-06-10 23:41:15 +07:00
Ali Maredia fc579f4520 ceph: initial commit for ceph rgw multisite resources
This commit contains CR implementations for:
CephObjectRealm
CephObjectZoneGroup
CephObjectZone

Also there are changes made to the objectstore
to add rgws in the object-store to zones and
zone groups in a multisite configuration and
the removal of the --default parameter for any
realms/zonegroups/zones that are created.

Signed-off-by: Ali Maredia <amaredia@redhat.com>
2020-06-09 16:28:53 -04:00
Madhu Rajanna 09ccf55453 ceph: update cephcsi to v2.1.2
Cephcsi v2.1.2 is released with bug fixes,
refer https://github.com/ceph/ceph-csi/releases/tag/v2.1.2
for release details.

Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
2020-06-05 14:43:05 +05:30
Satoru Takeuchi 91e7912336 ceph: fix the document of OSD management
Removing a disk is not mandatory after the data in the corresponding
OSD is moved to other OSDs. It's better to just let users to know it's
safe to remove the disk.

Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
2020-06-04 16:32:05 +00:00
Travis Nielsen c1ee1104c8 docs: fix indentation on the commit prefixes
The list of commit prefixes in the doc were rendering on
a single line instead of as a list.

Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
2020-05-28 17:02:12 -06:00
Travis Nielsen 0d1c83c9c3 docs: update the obc provisioner name
The OBC provisioner name changed recently to include the namespace
as a prefix, but the docs were not updated.

Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
2020-05-28 17:01:17 -06:00
Sébastien Han ff90fa7098 Merge pull request #5549 from leseb/semantic
add commitlint github action
2020-05-28 17:41:30 +02:00
Sébastien Han dea423fedd Merge pull request #5557 from cybozu-go/ceph-add-how-to-know-whether-filesystem-exist-on-top-of-device
ceph: add the description how to know whether there is a filesystem on top of a device
2020-05-28 17:38:11 +02:00
Sébastien Han db57455ba2 docs: update contribution guideline
Reflect the recent changes with the replacement of the commitlint bot
by the commitlint github action.

Signed-off-by: Sébastien Han <seb@redhat.com>
2020-05-28 17:36:08 +02:00
Satoru Takeuchi c3a2677883 ceph: add the description how to know whether there is a filesystem on top of a device
It's better to tell users to know whether their devices can be used as OSD
in the quick start guide.

Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
2020-05-28 15:06:57 +00:00
Sébastien Han c799e11032 ceph: enhancement cleanup with disks
Now, we can not only cleanup monitor data, logs and crashes but the
disks too. As part of the cleanupPolicy CR spec, we have a new setting
called sanitizeDisks which holds more details:

* confirmation: the confirmation message to sanitize disks,
use "yes-really-sanitize-disks" to confirm

This will **only** wipe the metadata, so it's a fast cleanup allowing
you to re-install later but won't remove all the data from the drive.

Signed-off-by: Sébastien Han <seb@redhat.com>
2020-05-27 18:14:02 +02:00
Sébastien Han a760b6d5ae doc: ceph fix rgw links
links were pointing to dead pages.

Signed-off-by: Sébastien Han <seb@redhat.com>
2020-05-13 17:16:49 +02:00
Satoru Takeuchi 75f0b0f159 ceph: write down the multipath support in the documents
OSD on PVC supports multipath thanks to the following commit.

ceph: add support for multipath devices
32730123ea

However, it's not documented yet.

Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
2020-05-12 17:11:02 +09:00
binoue ff2ba9c890 ceph: update ceph-upgrade.md for handling a known issue
Because of the ClusterRole api version change, #5341 may occur
when upgrading the rook chart to the latest version.
This pull request updates `ceph-upgrade.md` for handling the issue.

Signed-off-by: binoue <banji-inoue@cybozu.co.jp>
2020-05-11 03:22:55 +09:00
Travis Nielsen f82a07d441 Merge pull request #5394 from Madhu-1/csi-prio
CSI: Set PriorityClassName for CSI pods
2020-05-06 07:08:54 -06:00
Madhu Rajanna ec3ab55d94 CSI: Set PriorityClassName for CSI provisioner pods
provide an option to set PriorityclassName on the
CSI provisioner pods to avoid pod eviction, as the CSI provisioner
pods a critical in system for storage. the default value
for the CSI provisioner pods is set to system-cluster-critical.

provisioner pods are less priority than plugin pods.

Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
2020-05-06 15:45:06 +05:30
Madhu Rajanna cc6ec58891 CSI: Set PriorityClassName for CSI plugin pods
provide an option to set Priorityclassname on the
CSI plugin pods to avoid pod eviction, as the CSI plugin pods are
critical in system for storage. the default value
for the CSI plugin pods is set to system-node-critical.
more info can be found at
https://kubernetes.io/docs/tasks/administer-cluster/guaranteed-scheduling-critical-addon-pods/

Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
2020-05-06 15:45:01 +05:30