Age | Commit message (Collapse) | Author | Files | Lines |
|
Current startup script of etcd checks whether all assumed other nodes
are already running, before proceeding. This check, however, also
includes checking localhost, but due to using headless service
statefulset pod DNS discovery, it doesnt succeed immediately.
In some deployments k8s DNS server may be laggy, thus failing startup
script to finish before liveness check. This patch fixes such failures
of 1 pod etcd clusters, and improves startup time for any size ones.
Signed-off-by: Konrad Bańka <k.banka@samsung.com>
Issue-ID: OOM-2668
Change-Id: I2f9263a0f4964b0a495631775d0cbbceef25e85b
|
|
This commit makes etcd template to use the new generator for repositories and
images.
Issue-ID: OOM-2364
Signed-off-by: Sylvain Desbureaux <sylvain.desbureaux@orange.com>
Change-Id: Ia28c072aeef6bc58af3ec0acd584753df1f9b07c
|
|
Change-Id: I8cb12dae07cc3984e7dcfc602afa4c2d07317e9a
Signed-off-by: Jakub Latusek <j.latusek@samsung.com>
Issue-ID: OOM-2562
|
|
Issue-ID: OOM-2481
Signed-off-by: Daniel Milaszkiewicz <daniel.milaszkiewicz@nokia.com>
Change-Id: I3d2e249bfd158575c67eaf5600cdae1921bd3b2e
|
|
ONAP is too big to be deployed using helm install so we need to
use a custom helm plugin helm deploy. This script deloys onap
component by component instead of deploying evrything at
once. Unfortunately this script also modifies the helm release by
appending component name to it.
As a result of this behavior our objects are called for example:
onap-mariadb-galera-mariadb-galera-0
instead of just being called onap-mariadb-galera-0.
This patch simplifies this naming convention by replacing all direct
usages of .Release.Name with common.release macro which strips the
component specific part from the release name.
Issue-ID: OOM-2275
Signed-off-by: Krzysztof Opasiak <k.opasiak@samsung.com>
Change-Id: Ia8cead50d305adb00eef666d0a1ace74479b5183
|
|
OOM has now templates in order to create the needed PVC, using:
* a PV with a specific class when using a common nfs mount path between
nodes (sames as today use) --> is the default behavior today
* or a storage class if we want to use dynamic PV.
On this case, we use (in order of priority):
- persistence.storageClassOverride if set on the chart
- global.persistence.storageClass if set globally
- persistence.storageClass if set on the chart
I've also aligned the PV creation of the different charts.
I've also aligned the PVC creation of the different charts.
I've removed unused mysql chart and (badly) used nfs-provisioner chart.
I've also make cassandra backup work with dynamic PV (but RWX only for
now).
Change-Id: I0ea3f8c7514ca648d94b6c682684c06b822bbe0a
Issue-ID: OOM-2229
Issue-ID: OOM-2228
Issue-ID: OOM-2227
Issue-ID: OOM-1227
Signed-off-by: Sylvain Desbureaux <sylvain.desbureaux@orange.com>
|
|
Etcd pod fails to start with file not found error.
The error is being caused because the pod is stuck
in an infinite loop as the readiness probe is not true.
Since, we are anyway checking the pod status, we removed
the readiness probe as it is not needed.
Bumped up version of etcd to 3.2.24
This fixes a known issue:
https://github.com/etcd-io/etcd/pull/4861
Issue-ID: MULTICLOUD-660
Change-Id: I815766b4a8f187d88bb2fcdb71e9d6e24b277d25
Signed-off-by: Kiran Kamineni <kiran.k.kamineni@intel.com>
|
|
The chart uses servicename in statefulset.yaml.
This should be fullname instead.
Issue-ID: MULTICLOUD-636
Change-Id: I0e523a5c0e2e2c432f3a159fcb8c9ddcd2240e68
Signed-off-by: Kiran Kamineni <kiran.k.kamineni@intel.com>
|
|
Based on the helm chart in github helm repo
https://github.com/helm/charts/tree/master/incubator/etcd
etcd required as datastore by k8s Plugin in
multi-cloud project.
Change-Id: Ie6c7d9fc394f2a08cd9d34dda6261da13ef005f3
Signed-off-by: Ritu Sood <ritu.sood@intel.com>
Issue-ID: OOM-1828
|