profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/jvassev/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.

jvassev/easyevents 1

Easy events for spring, like in CDI

jvassev/git-objectstore-helper 1

Use S3 as a git repository using this remote-helper plugin for git

jvassev/ansible-zookeeper 0

Ansible playbook for ZooKeeper

jvassev/asdf-maven 0

Maven plugin for the asdf version manager.

jvassev/awesome-ipfs 0

Useful resources for using IPFS and building things on top of it

jvassev/aws-encryption-provider 0

APIServer encryption provider, backed by AWS KMS

jvassev/axios 0

Promise based HTTP client for the browser and node.js

jvassev/caffeine 0

A high performance caching library for Java 8

jvassev/charts 0

Curated applications for Kubernetes

issue openedkubernetes-client/javascript

Watch API breaks on IPv6

In our cluster kubernetes configured to use IPv6, and our service uses the K8S API internally from the cluster.

However starting a watch operation throws an error: Error [ERR_TLS_CERT_ALTNAME_INVALID]: Hostname/IP does not match certificate's altnames: Host: 2001. is not in the cert's altnames: DNS:kubernetes, DNS:kubernetes.default, DNS:kubernetes.default.svc, DNS:kubernetes.default.svc.cluster.local, IP Address:2001:1B70:82B9:18:0:0:2:1, IP Address:FD08:0:0:0:0:0:0:6, IP Address:2001:1B70:82B9:18:0:0:3:B, IP Address:2001:1B70:82B9:18:0:0:3:1, IP Address:FD08:0:0:0:0:0:0:7

After digging into the issue it turned out that the Request library which handles ipv6 hostnames wrongly: https://github.com/request/request/issues/3274

As Request is deprecated this issue likely won't be fixed.

Also I found that in the official latest K8S doc the proposed way to access the API to use DNS names: https://kubernetes.io/docs/tasks/administer-cluster/access-cluster-api/#directly-accessing-the-rest-api

The KUBERNETES_SERVICE_HOST and KUBERNETES_SERVICE_PORT is hardly mentioned, so I guess it is some deprecated way to do it.

As workaround adding a KUBERNETES_SERVICE_HOST=kubernetes.default.svc env parameter in the deployment config helped for us, but it may be a permanent solution: change the config.loadFromCluster() method and use well-known DNS name instead env variables, by default.

created time in 2 hours

pull request commentvmware/kube-fluentd-operator

feat(splunk-hec): add plugin & fix dependencies

@jliao2011 a new version of fluent-plugin-logzio (0.0.21) is available, please could you update your PR accordingly?

Done.

jliao2011

comment created time in 3 hours

issue commentkubernetes-client/javascript

Documentation not available

I also ran into this issue and have raised it with type-fest https://github.com/sindresorhus/type-fest/issues/184

clovis-dugue

comment created time in 8 hours

pull request commentvmware/kube-fluentd-operator

feat(splunk-hec): add plugin & fix dependencies

@jliao2011 a new version of fluent-plugin-logzio (0.0.21) is available, please could you update your PR accordingly?

jliao2011

comment created time in 9 hours

pull request commentvmware/kube-fluentd-operator

feat(splunk-hec): add plugin & fix dependencies

@jliao2011 Thanks for your contribution, there are teams using the Logz.io plugin in production so we cannot just remove this entirely. Looks like Logz.io are working on an update - https://github.com/logzio/fluent-plugin-logzio/issues/33

jliao2011

comment created time in a day

Pull request review commentvmware/kube-fluentd-operator

feat(splunk-hec): add plugin & fix dependencies

 gem 'fluent-plugin-amqp', "0.14.0" gem 'fluent-plugin-azure-loganalytics', "0.7.0" gem 'fluent-plugin-cloudwatch-logs', "0.8.0"

@jliao2011 Can we also update fluent-plugin-cloudwatch-logs to latest version v0.13.4 ?

jliao2011

comment created time in a day

push eventchris-armstrong/dynaglue

Chris Armstrong

commit sha 8875c46c1c6694f4ed13fd220a981094c751d0e9

upgrade typedoc

view details

Chris Armstrong

commit sha 444ff4aa6aa3d63a4cda59e17be588df261bb090

updated docs

view details

push time in 2 days

PR opened jeremydaly/dynamodb-toolbox

Update README.md

I just added a few details on startKey and secondary indexes that threw me for a while - thought others might find it useful!

+1 -1

0 comment

1 changed file

pr created time in 2 days

push eventkubevault/csi-driver

1gtm

commit sha cde17441804c27dabaa7a3609b25e6d61b1ff29c

Update Kubernetes v1.18.9 dependencies Signed-off-by: 1gtm <1gtm@appscode.com>

view details

push time in 3 days

issue openedkubernetes-client/javascript

Issue with `exec` command

I'm trying to run the following code and I can't get the $(which bundle) to work. Any suggestions?

  const cluster = getClusterClient(env, region)
  const exec = new k8s.Exec(cluster)
  const rl = await getReadLine()
  try {
    await exec.exec(
      pod.metadata.namespace,
      pod.metadata.name,
      pod.spec.containers[0].name,
      ['/bin/bash', '-c', 'export BUNDLE_CMD=$(which bundle); echo $BUNDLE_CMD'],
      rl.output, process.stderr, rl.input, true)
  } catch (error) {
    logger.error(error)
    rl.close()
  }

created time in 3 days

push eventkubevault/csi-driver

1gtm

commit sha e8773c1d7117173793c0e7e83a9e928066401c64

Update Kubernetes v1.18.9 dependencies Signed-off-by: 1gtm <1gtm@appscode.com>

view details

push time in 4 days

issue commentkubernetes-client/javascript

Option to specify Resync period in Informers

But why do you need the resync? If your controller needs a complete list it can call the list method at any time? When you create a new informer you'll get a complete list of all objects.

Why does your controller need a periodic update event for all objects?

thejasbabu

comment created time in 4 days

issue commentvmware/kube-fluentd-operator

No solution to configure the log level of the vmware/kube-fluentd-operator

Yes that would be nice to propagate the log-level to both containers 👍

ls-julien-deruere

comment created time in 4 days

push eventkubevault/csi-driver

1gtm

commit sha 4cae51261f5814845127bf7fdd2750eea46d49bb

Update Kubernetes v1.18.9 dependencies Signed-off-by: 1gtm <1gtm@appscode.com>

view details

push time in 5 days

issue commentkubernetes-client/javascript

Option to specify Resync period in Informers

No, I was thinking more in terms of an informer which internally manages the resync mechanisms. Similar to what the client-go provides: https://github.com/kubernetes/client-go/blob/8f8241f9ef2dc939ecaf3cbdc80f6eb58227efd9/tools/cache/controller.go#L344

thejasbabu

comment created time in 5 days

issue commentkubernetes-client/javascript

Option to specify Resync period in Informers

The informer also implements the ListWatch interface:

Which has a list() method which will always give you the complete list of objects. https://github.com/kubernetes-client/javascript/blob/master/src/cache.ts#L72

Does that accomplish what you need?

thejasbabu

comment created time in 5 days

issue commentkubernetes-client/javascript

watcher connection stops receiving events after some time

In general, I think relying on watches to stick around for greater than N minutes (where N is fairly small) is probably not the best bet. I think adding a timeout if an event hasn't been received in N minutes (you can tune N depending on your workload) is probably the right approach.

The network is just too weird a place to expect that HTTP/TCP connections will stay alive for a long time.

jkryl

comment created time in 5 days

issue commentkubernetes-client/javascript

Documentation not available

There's also the Kubernetes API reference which applies to the resource objects:

https://kubernetes.io/docs/reference/

clovis-dugue

comment created time in 5 days

issue commentkubernetes-client/javascript

Documentation not available

I find that GitHub search is adequate. Also, there are lots of examples here:

https://github.com/kubernetes-client/javascript/tree/master/examples

clovis-dugue

comment created time in 5 days

pull request commentkubernetes-client/javascript

cache: skip callback upon restart if object did not change

Thanks! One comment wrt null checking

jkryl

comment created time in 5 days

Pull request review commentkubernetes-client/javascript

cache: skip callback upon restart if object did not change

 function isSameObject<T extends KubernetesObject>(o1: T, o2: T): boolean {     return o1.metadata!.name === o2.metadata!.name && o1.metadata!.namespace === o2.metadata!.namespace; } +function isSameVersion<T extends KubernetesObject>(o1: T, o2: T): boolean {+    return (+        o1.metadata!.resourceVersion !== undefined &&

probably worth testing for null also?

jkryl

comment created time in 5 days

issue commentvmware/kube-fluentd-operator

No solution to configure the log level of the vmware/kube-fluentd-operator

Edit: Sorry, yes. We can figure out a quick way to propogate log-level in the generator.go code: https://github.com/vmware/kube-fluentd-operator/blob/master/config-reloader/generator/generator.go

ls-julien-deruere

comment created time in 5 days

issue commentvmware/kube-fluentd-operator

No solution to configure the log level of the vmware/kube-fluentd-operator

Yes, config-reloader has it's own log-level that is independently configured outside of fluentd. As config-reloader is written in Go and verifies/reloads the configs for fluentd.

Have you tried setting the log-level field in your fluentd.conf file: https://github.com/vmware/kube-fluentd-operator/blob/master/config-reloader/templates/fluent.conf#L4

This can be done through the configMap that fluent-operator uses: https://github.com/vmware/kube-fluentd-operator/blob/master/config-reloader/examples/manifests/kfo-test.yaml#L142

Possible values are: Specifies the log level e.g. trace, debug, info, warn, error, or fatal (https://docs.fluentd.org/deployment/system-config)

ls-julien-deruere

comment created time in 5 days

issue commentvmware/kube-fluentd-operator

Config-reloader hangs during config check

We have recently merged in some timeouts and log fixes (https://github.com/vmware/kube-fluentd-operator/pull/180) in order to make config-reloader a bit more resilient and fail quicker in such cases where it cannot validate the fluentd configurations. It will fail quicker and log out more verbose messages.

We will be making a new release soon.

a-b-v

comment created time in 5 days

issue openedvmware/kube-fluentd-operator

No solution to configure the log level of the vmware/kube-fluentd-operator

I'm using the Helm chart, the the logLevel only apply the the config-reloader. But the Fluentd container has no way to configure the log level.

I end up having a lot of noise: level=debug msg="Using default configmap name ('fluentd-config') for namespace 'kube-system'"

created time in 5 days

issue commentvmware/kube-fluentd-operator

Config-reloader hangs during config check

No I've migrated to other solution.

a-b-v

comment created time in 5 days

issue commentvmware/kube-fluentd-operator

Config-reloader hangs during config check

Hey, I'm having exactly same issue. Have you found a workaround? I'm afraid it's going to hang on production cluster and https://github.com/vmware/kube-fluentd-operator/blob/master/charts/log-router/templates/daemonset.yaml#L100

a-b-v

comment created time in 5 days

pull request commentkubernetes-client/javascript

cache: skip callback upon restart if object did not change

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: <a href="https://github.com/kubernetes-client/javascript/pull/597#" title="Author self-approved">jkryl</a> To complete the pull request process, please assign brendandburns after the PR has been reviewed. You can assign the PR to them by writing /assign @brendandburns in a comment when ready.

The full list of commands accepted by this bot can be found here.

<details open> Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment Approvers can cancel approval by writing /approve cancel in a comment </details> <!-- META={"approvers":["brendandburns"]} -->

jkryl

comment created time in 5 days

PR opened kubernetes-client/javascript

cache: skip callback upon restart if object did not change

Currently MOD callbacks are called for all objects in the cache when the watcher is restarted. This commit makes the algorithm smarter and call MOD callback only if the object has really changed. The decision is based on "resourceVersion" metadata field in the object.

+19 -4

0 comment

2 changed files

pr created time in 5 days

issue commentkubernetes-client/javascript

watcher connection stops receiving events after some time

let me correct myself. The keep-alives shown by tcpdump were not keep-alives sent by the watcher but they are sent by the api server every 3 minutes:

...
11:32:59.239617 IP 227.114.246.35.bc.googleusercontent.com.https > gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832: Flags [.], ack 1, win 501, options [nop,nop,TS val 2612916151 ecr 3766451319], length 0
11:32:59.239753 IP gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832 > 227.114.246.35.bc.googleusercontent.com.https: Flags [.], ack 19557, win 331, options [nop,nop,TS val 3766631543 ecr 2611652359], length 0
11:35:59.463662 IP 227.114.246.35.bc.googleusercontent.com.https > gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832: Flags [.], ack 1, win 501, options [nop,nop,TS val 2613096375 ecr 3766631543], length 0
11:35:59.463762 IP gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832 > 227.114.246.35.bc.googleusercontent.com.https: Flags [.], ack 19557, win 331, options [nop,nop,TS val 3766811767 ecr 2611652359], length 0

Now comes the breakpoint roughly after half an hour since the connection has been established. The watcher receives a couple of bytes from the server and immediately starts to send keep-alives every second. That's when the connection becomes dead and does not receive any new events.

11:38:09.258810 IP 227.114.246.35.bc.googleusercontent.com.https > gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832: Flags [P.], seq 19557:19584, ack 1, win 501, options [nop,nop,TS val 2613226170 ecr 3766811767], length 27
11:38:09.258890 IP gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832 > 227.114.246.35.bc.googleusercontent.com.https: Flags [.], ack 19584, win 331, options [nop,nop,TS val 3766941562 ecr 2613226170], length 0
11:38:10.273184 IP gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832 > 227.114.246.35.bc.googleusercontent.com.https: Flags [.], ack 19584, win 331, options [nop,nop,TS val 3766942576 ecr 2613226170], length 0
11:38:10.273612 IP 227.114.246.35.bc.googleusercontent.com.https > gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832: Flags [.], ack 1, win 501, options [nop,nop,TS val 2613227185 ecr 3766941562], length 0
11:38:11.297176 IP gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832 > 227.114.246.35.bc.googleusercontent.com.https: Flags [.], ack 19584, win 331, options [nop,nop,TS val 3766943600 ecr 2613227185], length 0
11:38:11.297533 IP 227.114.246.35.bc.googleusercontent.com.https > gke-jan-test2-default-pool-664e7812-gs7g.c.mayastor.internal.55832: Flags [.], ack 1, win 501, options [nop,nop,TS val 2613228209 ecr 3766941562], length 0
...

My keep-alive patch does not seem to change this behaviour. Either it's not the way to go or it is not complete and still missing something. For the reference this is the commit I have been testing: https://github.com/jkryl/javascript/commit/0fef4de1e7c1dacfa7da04f1e4f6e2047525c9f3#diff-8af1000c89b03ced37f439c61c5696c45e1e83a70cc07182feef6595123f0bad

jkryl

comment created time in 5 days