upgrade: retry if default DSCI creation fails #1008
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
/retest |
1 similar comment
|
/retest |
zdtsw
approved these changes
May 14, 2024
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: AjayJagan, zdtsw The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
/retest |
After removing leader election, operator fails to start if it is
instructed to create default DSCI. Looks like webhook is not ready
by the time:
```
create default DSCI CR.
{"level":"error","ts":"2024-05-13T09:25:58Z","logger":"setup","msg":"unable to create initial setup for the operator","error":"Internal error occurred: failed calling webhook \"operator.opendatahub.io\": failed to call webhook: Post \"https://opendatahub-operator-controller-manager-service.oo-2ts9m.svc:443/validate-opendatahub-io-v1?timeout=10s\": no endpoints available for service \"opendatahub-operator-controller-manager-service\"","stacktrace":"main.main.func1\n\t/workspace/main.go:200\nsigs.k8s.io/controller-runtime/pkg/manager.RunnableFunc.Start\n\t/remote-source/operator/deps/gomod/pkg/mod/sigs.k8s.io/controller-runtime@v0.14.6/pkg/manager/manager.go:336\nsigs.k8s.io/controller-runtime/pkg/manager.(*runnableGroup).reconcile.func1\n\t/remote-source/operator/deps/gomod/pkg/mod/sigs.k8s.io/controller-runtime@v0.14.6/pkg/manager/runnable_group.go:219"}
```
Leader election added some delay.
The problem does not happen in default configuration since it
explicitly disables DSCI creation in the manifests:
```
containers:
- command:
- /manager
env:
- name: DISABLE_DSC_CONFIG
value: 'true'
args:
- --operator-name=opendatahub
image: controller:latest
```
Make a wrapper function cluster.CreateWithRetry for client.Object
creation with timeout. Use hardcoded 5s interval, just seems
reasonable, and timeout in minutes as the parameter.
It requires disable linter nilerr since for the polling function
error in creation is a valid condition, something the function wait
to disappear.
Fixes: 3610b0b ("feat: remove leader election for operator (opendatahub-io#1000)")
Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
a7889fb to
a892b10
Compare
|
Actually, generic is overthinking, taking interface here is enough. |
|
/retest |
|
/retest-required |
|
/lgtm |
e26100e
into
opendatahub-io:incubation
8 checks passed
VaishnaviHire
added a commit
that referenced
this pull request
May 14, 2024
* Update version to v2.12.0 (#1007) * upgrade: retry if default DSCI creation fails (#1008) After removing leader election, operator fails to start if it is instructed to create default DSCI. Looks like webhook is not ready by the time: ``` create default DSCI CR. {"level":"error","ts":"2024-05-13T09:25:58Z","logger":"setup","msg":"unable to create initial setup for the operator","error":"Internal error occurred: failed calling webhook \"operator.opendatahub.io\": failed to call webhook: Post \"https://opendatahub-operator-controller-manager-service.oo-2ts9m.svc:443/validate-opendatahub-io-v1?timeout=10s\": no endpoints available for service \"opendatahub-operator-controller-manager-service\"","stacktrace":"main.main.func1\n\t/workspace/main.go:200\nsigs.k8s.io/controller-runtime/pkg/manager.RunnableFunc.Start\n\t/remote-source/operator/deps/gomod/pkg/mod/sigs.k8s.io/controller-runtime@v0.14.6/pkg/manager/manager.go:336\nsigs.k8s.io/controller-runtime/pkg/manager.(*runnableGroup).reconcile.func1\n\t/remote-source/operator/deps/gomod/pkg/mod/sigs.k8s.io/controller-runtime@v0.14.6/pkg/manager/runnable_group.go:219"} ``` Leader election added some delay. The problem does not happen in default configuration since it explicitly disables DSCI creation in the manifests: ``` containers: - command: - /manager env: - name: DISABLE_DSC_CONFIG value: 'true' args: - --operator-name=opendatahub image: controller:latest ``` Make a wrapper function cluster.CreateWithRetry for client.Object creation with timeout. Use hardcoded 5s interval, just seems reasonable, and timeout in minutes as the parameter. It requires disable linter nilerr since for the polling function error in creation is a valid condition, something the function wait to disappear. Fixes: 3610b0b ("feat: remove leader election for operator (#1000)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> --------- Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> Co-authored-by: Yauheni Kaliuta <ykaliuta@redhat.com>
ykaliuta
added a commit
to ykaliuta/opendatahub-operator
that referenced
this pull request
Jun 6, 2024
DSC creation is checked by the webhook so will not work from the main before manager starts if webhook is enabled. The same as DSCI. It also requires CreateWithRetry() for the same reasons as e26100e ("upgrade: retry if default DSCI creation fails (opendatahub-io#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
ykaliuta
added a commit
to ykaliuta/opendatahub-operator
that referenced
this pull request
Jun 7, 2024
DSC creation is checked by the webhook so will not work from the main before manager starts if webhook is enabled. The same as DSCI. It also requires CreateWithRetry() for the same reasons as e26100e ("upgrade: retry if default DSCI creation fails (opendatahub-io#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
openshift-merge-bot bot
pushed a commit
that referenced
this pull request
Jun 12, 2024
* main: move DSC creation after DSCIs one Reorder a bit to keep definition of cleanExistingResourceFunc next to the Add() call. Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> * main: convert DSC creation to RunableFunc DSC creation is checked by the webhook so will not work from the main before manager starts if webhook is enabled. The same as DSCI. It also requires CreateWithRetry() for the same reasons as e26100e ("upgrade: retry if default DSCI creation fails (#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> --------- Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
zdtsw
referenced
this pull request
in zdtsw-forking/rhods-operator
Jun 18, 2024
…1008) After removing leader election, operator fails to start if it is instructed to create default DSCI. Looks like webhook is not ready by the time: ``` create default DSCI CR. {"level":"error","ts":"2024-05-13T09:25:58Z","logger":"setup","msg":"unable to create initial setup for the operator","error":"Internal error occurred: failed calling webhook \"operator.opendatahub.io\": failed to call webhook: Post \"https://opendatahub-operator-controller-manager-service.oo-2ts9m.svc:443/validate-opendatahub-io-v1?timeout=10s\": no endpoints available for service \"opendatahub-operator-controller-manager-service\"","stacktrace":"main.main.func1\n\t/workspace/main.go:200\nsigs.k8s.io/controller-runtime/pkg/manager.RunnableFunc.Start\n\t/remote-source/operator/deps/gomod/pkg/mod/sigs.k8s.io/controller-runtime@v0.14.6/pkg/manager/manager.go:336\nsigs.k8s.io/controller-runtime/pkg/manager.(*runnableGroup).reconcile.func1\n\t/remote-source/operator/deps/gomod/pkg/mod/sigs.k8s.io/controller-runtime@v0.14.6/pkg/manager/runnable_group.go:219"} ``` Leader election added some delay. The problem does not happen in default configuration since it explicitly disables DSCI creation in the manifests: ``` containers: - command: - /manager env: - name: DISABLE_DSC_CONFIG value: 'true' args: - --operator-name=opendatahub image: controller:latest ``` Make a wrapper function cluster.CreateWithRetry for client.Object creation with timeout. Use hardcoded 5s interval, just seems reasonable, and timeout in minutes as the parameter. It requires disable linter nilerr since for the polling function error in creation is a valid condition, something the function wait to disappear. Fixes: 3610b0b ("feat: remove leader election for operator (red-hat-data-services#1000)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
zdtsw
referenced
this pull request
in zdtsw-forking/rhods-operator
Jun 18, 2024
* main: move DSC creation after DSCIs one Reorder a bit to keep definition of cleanExistingResourceFunc next to the Add() call. Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> * main: convert DSC creation to RunableFunc DSC creation is checked by the webhook so will not work from the main before manager starts if webhook is enabled. The same as DSCI. It also requires CreateWithRetry() for the same reasons as e26100e ("upgrade: retry if default DSCI creation fails (red-hat-data-services#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> --------- Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
VaishnaviHire
pushed a commit
to VaishnaviHire/opendatahub-operator
that referenced
this pull request
Jun 18, 2024
* main: move DSC creation after DSCIs one Reorder a bit to keep definition of cleanExistingResourceFunc next to the Add() call. Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> * main: convert DSC creation to RunableFunc DSC creation is checked by the webhook so will not work from the main before manager starts if webhook is enabled. The same as DSCI. It also requires CreateWithRetry() for the same reasons as e26100e ("upgrade: retry if default DSCI creation fails (opendatahub-io#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> --------- Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> (cherry picked from commit 835959a)
ykaliuta
added a commit
to ykaliuta/opendatahub-operator
that referenced
this pull request
Jul 30, 2024
Intention of the function was to ensure that the object is created
and retry in case of webhook is temporary not available.
The status error looks like:
```
error(*k8s.io/apimachinery/pkg/api/errors.StatusError) *{
ErrStatus: k8s.io/apimachinery/pkg/apis/meta/v1.Status {
TypeMeta: (*"k8s.io/apimachinery/pkg/apis/meta/v1.TypeMeta")(0xc000504500),
ListMeta: (*"k8s.io/apimachinery/pkg/apis/meta/v1.ListMeta")(0xc000504520),
Status: "Failure",
Message: "Internal error occurred: failed calling webhook \"operator.openda...+194 more",
Reason: "InternalError",
Details: *(*"k8s.io/apimachinery/pkg/apis/meta/v1.StatusDetails")(0xc00046c9c0),
Code: 500,},}
```
which makes filtering of it tricky, other real errors can be in
theory reported as "InternalError" code 500. So, the original code
considered any error as a retry case.
Unfortunately, it made AlreadyExists as a retry condition which is
acutally the real goal of the function (make the object exists).
Ignore the error in error check making it return successfully in
that case.
Fixes: e26100e ("upgrade: retry if default DSCI creation fails (opendatahub-io#1008)")
Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
5 tasks
ykaliuta
added a commit
to ykaliuta/opendatahub-operator
that referenced
this pull request
Jul 30, 2024
Intention of the function was to ensure that the object is created
and retry in case of webhook is temporary not available.
The status error looks like:
```
error(*k8s.io/apimachinery/pkg/api/errors.StatusError) *{
ErrStatus: k8s.io/apimachinery/pkg/apis/meta/v1.Status {
TypeMeta: (*"k8s.io/apimachinery/pkg/apis/meta/v1.TypeMeta")(0xc000504500),
ListMeta: (*"k8s.io/apimachinery/pkg/apis/meta/v1.ListMeta")(0xc000504520),
Status: "Failure",
Message: "Internal error occurred: failed calling webhook \"operator.openda...+194 more",
Reason: "InternalError",
Details: *(*"k8s.io/apimachinery/pkg/apis/meta/v1.StatusDetails")(0xc00046c9c0),
Code: 500,},}
```
which makes filtering of it tricky, other real errors can be in
theory reported as "InternalError" code 500. So, the original code
considered any error as a retry case.
Unfortunately, it made AlreadyExists as a retry condition which is
acutally the real goal of the function (make the object exists).
Ignore the error in error check making it return successfully in
that case.
Fixes: e26100e ("upgrade: retry if default DSCI creation fails (opendatahub-io#1008)")
Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
ykaliuta
added a commit
to ykaliuta/opendatahub-operator
that referenced
this pull request
Jul 30, 2024
Intention of the function was to ensure that the object is created and retry in case of webhook is temporary not available. The original code lacks handling already existing object. Handling with and without webhook configurations make the things more complicated. Let's check Create() errors for different cases: If webhook enabled: - no error (err == nil) - 500 InternalError likely if webhook is not available (yet) - 403 Forbidden if webhook blocks creation (check of existance) - some problem (real error) else, if webhook disabled: - no error (err == nil) - 409 AlreadyExists if object exists - some problem (real error) Check already existing object after Create() with a call to Get(). It covers both with and without webhook configurations. Reuse the same object for Get() to avoid fetching Gvk for it. It doesn't make harm, the object structure is not used after creation. 500 InternalError is not 100% webhook problem, but consider it as a reason for retry. Fixes: e26100e ("upgrade: retry if default DSCI creation fails (opendatahub-io#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
ykaliuta
added a commit
to ykaliuta/opendatahub-operator
that referenced
this pull request
Aug 1, 2024
Intention of the function was to ensure that the object is created and retry in case of webhook is temporary not available. The original code lacks handling already existing object. Handling with and without webhook configurations make the things more complicated. Let's check Create() errors for different cases: If webhook enabled: - no error (err == nil) - 500 InternalError likely if webhook is not available (yet) - 403 Forbidden if webhook blocks creation (check of existance) - some problem (real error) else, if webhook disabled: - no error (err == nil) - 409 AlreadyExists if object exists - some problem (real error) Check already existing object after Create() with a call to Get(). It covers both with and without webhook configurations. Reuse the same object for Get() to avoid fetching Gvk for it. It doesn't make harm, the object structure is not used after creation. 500 InternalError is not 100% webhook problem, but consider it as a reason for retry. Fixes: e26100e ("upgrade: retry if default DSCI creation fails (opendatahub-io#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
ykaliuta
added a commit
to ykaliuta/opendatahub-operator
that referenced
this pull request
Aug 5, 2024
Intention of the function was to ensure that the object is created and retry in case of webhook is temporary not available. The original code lacks handling already existing object. Handling with and without webhook configurations make the things more complicated. Let's check Create() errors for different cases: If webhook enabled: - no error (err == nil) - 500 InternalError likely if webhook is not available (yet) - 403 Forbidden if webhook blocks creation (check of existance) - some problem (real error) else, if webhook disabled: - no error (err == nil) - 409 AlreadyExists if object exists - some problem (real error) Check already existing object after Create() with a call to Get(). It covers both with and without webhook configurations. Reuse the same object for Get() to avoid fetching Gvk for it. It doesn't make harm, the object structure is not used after creation. 500 InternalError is not 100% webhook problem, but consider it as a reason for retry. Fixes: e26100e ("upgrade: retry if default DSCI creation fails (opendatahub-io#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
openshift-merge-bot bot
pushed a commit
that referenced
this pull request
Aug 5, 2024
* cluster: CreateWithRetry: make work for already existing object Intention of the function was to ensure that the object is created and retry in case of webhook is temporary not available. The original code lacks handling already existing object. Handling with and without webhook configurations make the things more complicated. Let's check Create() errors for different cases: If webhook enabled: - no error (err == nil) - 500 InternalError likely if webhook is not available (yet) - 403 Forbidden if webhook blocks creation (check of existance) - some problem (real error) else, if webhook disabled: - no error (err == nil) - 409 AlreadyExists if object exists - some problem (real error) Check already existing object after Create() with a call to Get(). It covers both with and without webhook configurations. Reuse the same object for Get() to avoid fetching Gvk for it. It doesn't make harm, the object structure is not used after creation. 500 InternalError is not 100% webhook problem, but consider it as a reason for retry. Fixes: e26100e ("upgrade: retry if default DSCI creation fails (#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> * upgrade: cleanup CreateWithRetry usage CreateWithRetry() now checks AlreadyExists condition inside, so skip its handling. sleep() is not needed for DSCI with proper working CreateWithRetry since it is the actual point of existance of the function. Keep checking number of DSCI instances for non-webhook configuration. Basically, using CreateWithRetry for DSC is redundant from webhook unavailability point of view since it is created after DSCI which garantees working webhook. But it handles already existed object. Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> --------- Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com>
zdtsw
referenced
this pull request
in zdtsw-forking/rhods-operator
Aug 15, 2024
* cluster: CreateWithRetry: make work for already existing object Intention of the function was to ensure that the object is created and retry in case of webhook is temporary not available. The original code lacks handling already existing object. Handling with and without webhook configurations make the things more complicated. Let's check Create() errors for different cases: If webhook enabled: - no error (err == nil) - 500 InternalError likely if webhook is not available (yet) - 403 Forbidden if webhook blocks creation (check of existance) - some problem (real error) else, if webhook disabled: - no error (err == nil) - 409 AlreadyExists if object exists - some problem (real error) Check already existing object after Create() with a call to Get(). It covers both with and without webhook configurations. Reuse the same object for Get() to avoid fetching Gvk for it. It doesn't make harm, the object structure is not used after creation. 500 InternalError is not 100% webhook problem, but consider it as a reason for retry. Fixes: e26100e ("upgrade: retry if default DSCI creation fails (red-hat-data-services#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> * upgrade: cleanup CreateWithRetry usage CreateWithRetry() now checks AlreadyExists condition inside, so skip its handling. sleep() is not needed for DSCI with proper working CreateWithRetry since it is the actual point of existance of the function. Keep checking number of DSCI instances for non-webhook configuration. Basically, using CreateWithRetry for DSC is redundant from webhook unavailability point of view since it is created after DSCI which garantees working webhook. But it handles already existed object. Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> --------- Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> (cherry picked from commit 5c755e6)
openshift-merge-bot bot
referenced
this pull request
in red-hat-data-services/rhods-operator
Aug 16, 2024
* cluster: CreateWithRetry: make work for already existing object Intention of the function was to ensure that the object is created and retry in case of webhook is temporary not available. The original code lacks handling already existing object. Handling with and without webhook configurations make the things more complicated. Let's check Create() errors for different cases: If webhook enabled: - no error (err == nil) - 500 InternalError likely if webhook is not available (yet) - 403 Forbidden if webhook blocks creation (check of existance) - some problem (real error) else, if webhook disabled: - no error (err == nil) - 409 AlreadyExists if object exists - some problem (real error) Check already existing object after Create() with a call to Get(). It covers both with and without webhook configurations. Reuse the same object for Get() to avoid fetching Gvk for it. It doesn't make harm, the object structure is not used after creation. 500 InternalError is not 100% webhook problem, but consider it as a reason for retry. Fixes: e26100e ("upgrade: retry if default DSCI creation fails (#1008)") Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> * upgrade: cleanup CreateWithRetry usage CreateWithRetry() now checks AlreadyExists condition inside, so skip its handling. sleep() is not needed for DSCI with proper working CreateWithRetry since it is the actual point of existance of the function. Keep checking number of DSCI instances for non-webhook configuration. Basically, using CreateWithRetry for DSC is redundant from webhook unavailability point of view since it is created after DSCI which garantees working webhook. But it handles already existed object. Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> --------- Signed-off-by: Yauheni Kaliuta <ykaliuta@redhat.com> (cherry picked from commit 5c755e6)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
After removing leader election, operator fails to start if it is instructed to create default DSCI. Looks like webhook is not ready by the time:
Leader election added some delay.
The problem does not happen in default configuration since it explicitly disables DSCI creation in the manifests:
Make a wrapper function cluster.CreateWithRetry for client.Object creation with timeout. Use 5s interval, just seems reasonable, and timeout in minutes as the parameter.
It requires disable linter nilerr since for the polling function error in creation is a valid condition, something the function wait to disappear.
Fixes: 3610b0b ("feat: remove leader election for operator (#1000)")
How Has This Been Tested?
Merge criteria: