profile
viewpoint
Karsten Jeschkies jeschkies http://www.blackmagiclabs.com/ Software engineer working @grafana .

jenkinsci/mesos-plugin 290

Mesos Cloud Jenkins Plugin

boxdot/letterboxd-rs 7

Letterboxd API for access to data on the Letterboxd.com website in Rust.

jeschkies/async-mesos-rs 3

An asynchronous client for Mesos HTTP scheduler API.

jeschkies/lox-rs 2

A Lox Interpreter in Rust

boxdot/basic-rs 1

Minimal BASIC interpreter written in Rust

jeschkies/activator-akka-distributed-workers 0

Typesafe Activator template for distributed workers with Akka cluster.

jeschkies/algorithms 0

A collection of little algorithms I implemented for practice and fun.

jeschkies/aphorisms 0

A List of Aphorims I Try to Follow in My Engineering Life

jeschkies/authd 0

An Authentication Daemon

startedslim-bean/loki-shell

started time in 24 days

Pull request review commentgrafana/agent

override remote_name when grouping configs

 func groupConfigs(groupName string, grouped groupedConfigs) (Config, error) { 	combined.Name = groupName 	combined.ScrapeConfigs = []*config.ScrapeConfig{} +	// Assign all remote_write configs in the group a consistent set of remote_names.+	// If the grouped configs are coming from the scraping service, defaults will have+	// been applied and the remote names will be prefixed with the old instance config name.+	for _, rwc := range combined.RemoteWrite {+		// Blank out the existing name before getting the hash so it is doesn't take into+		// account any existing name.+		rwc.Name = ""++		hash, err := getHash(rwc)+		if err != nil {+			return Config{}, err+		}++		rwc.Name = groupName[:6] + "-" + hash[:6]

Out of curiosity, why the first six?

rfratto

comment created time in 25 days

Pull request review commentgrafana/agent

override remote_name when grouping configs

   config API endpoint to prevent a race condition of two conflicting configs   being applied at the same time. (@rfratto) +- [BUGFIX] remote_write names in a group will no longer be copied from the
- [BUGFIX] `remote_write` names in a group will no longer be copied from the

just nit picking :slightly_smiling_face:

rfratto

comment created time in 25 days

PullRequestReviewEvent
PullRequestReviewEvent

delete branch jeschkies/cortex

delete branch : karsten/mysql-e2e

delete time in a month

create barnchjeschkies/cortex

branch : karsten/mysql-e2e

created branch time in a month

fork jeschkies/cortex

A horizontally scalable, highly available, multi-tenant, long term Prometheus.

https://cortexmetrics.io/

fork in a month

push eventjenkinsci/mesos-plugin

awltux

commit sha a4a7197c4729b98d52f4cfa68a9bee04a327b9a8

Jenkins throws exceptions showing configuration for a Mesos Agent (#407) (#409) Co-authored-by: Chris Holman <Chris.Holman@eu.medical.canon>

view details

push time in a month

PR merged jenkinsci/mesos-plugin

Jenkins throws exceptions showing configuration for a Mesos Agent (#407)

A simple fix for a minor issue; but I don't like seeing error pages.

+14 -0

0 comment

2 changed files

awltux

pr closed time in a month

push eventjeschkies/dotfiles

Karsten Jeschkies

commit sha 3f0d70accea8c04b0c6a1ddeaa3ae56c6c1cf517

Load fzf key bindings last.

view details

push time in a month

push eventjeschkies/dotfiles

Karsten Jeschkies

commit sha 47f1f2b045aecaa7dda5c275121376128e9262e1

Import key bindings for fzf.

view details

push time in a month

push eventjeschkies/dotfiles

Karsten Jeschkies

commit sha d024b1f3dbf413419bf91199f81536a05fbcb3d5

Configure VIM again.

view details

push time in a month

push eventjeschkies/dotfiles

Karsten Jeschkies

commit sha 96daabd0997a1f4154243d53b945eade5640eda2

Fedora work station.

view details

Karsten Jeschkies

commit sha 967a3b8f0ae86eb8e70ca1605e55c7d244b88cee

Merge branch 'master' into karsten

view details

Karsten Jeschkies

commit sha 855e70bdad8c419859176dce582c52bed330ad80

Remove `hub`

view details

push time in a month

delete branch grafana/grafonnet-lib

delete branch : karsten/bar-gauge

delete time in 2 months

push eventgrafana/grafonnet-lib

Karsten Jeschkies

commit sha 8338be68387b5811355aa919d031882ead0c9c6e

Define bar gauge. (#249) Summary: This adds a simple bar gauge panel builder. Only title, datasource and unit can be defined.

view details

push time in 2 months

PR merged grafana/grafonnet-lib

Reviewers
Define bar gauge.

Summary: This adds a simple bar gauge panel builder. Only title, datasource and unit can be defined.

+48 -0

3 comments

2 changed files

jeschkies

pr closed time in 2 months

pull request commentgrafana/grafonnet-lib

Define bar gauge.

Friendly ping @trotttrotttrott

jeschkies

comment created time in 2 months

push eventjeschkies/gurk-rs

Karsten Jeschkies

commit sha a3ca019ecc7ccfe784ef4b34e8ada1bc14c99951

Load channels.

view details

push time in 2 months

push eventjeschkies/gurk-rs

Karsten Jeschkies

commit sha f80a6970be8f66ee4ebc82e1b2d568cbf0f35899

Save subset of messages and channels.

view details

push time in 2 months

push eventgrafana/grafonnet-lib

Karsten Jeschkies

commit sha 26ef484f608e1efb95964f0ec49ac8d22af06eb2

fixup! make description and thresholds configurable.

view details

push time in 2 months

pull request commentgrafana/grafonnet-lib

Define bar gauge.

I see the sibling gauge_panel.libsonnet has a whole lot of parameters, but I imagine any extra ones could just be added as needed?

Yes, there are a lot. I would add them as we need them. At least that's my proposal.

jeschkies

comment created time in 2 months

push eventgrafana/grafonnet-lib

Karsten Jeschkies

commit sha 446edb6251266b5fd85282ef8db52cd18637bb37

fixup! format code.

view details

push time in 2 months

Pull request review commentgrafana/grafonnet-lib

Define bar gauge.

+{

It is not formatted well but I did not find a way how to install the formatter.

jeschkies

comment created time in 2 months

PR opened grafana/grafonnet-lib

Reviewers
Define bar gauge.

Summary: This adds a simple bar gauge panel builder. Only title, datasource and unit can be defined.

+39 -0

0 comment

1 changed file

pr created time in 2 months

create barnchgrafana/grafonnet-lib

branch : karsten/bar-gauge

created branch time in 2 months

Pull request review commentgrafana/agent

allow dynamic updating of instance scrape_configs and remote_write

 func (i *Instance) getRemoteWriteTimestamp() int64 { 	return ts * 1000 } +// Config returns the current Config stored by the Instance.+func (i *Instance) Config() Config {+	i.cfgMutex.Lock()+	defer i.cfgMutex.Unlock()+	return i.cfg

Ah, got it. Thanks :slightly_smiling_face:

rfratto

comment created time in 2 months

Pull request review commentgrafana/agent

allow dynamic updating of instance scrape_configs and remote_write

 func (i *Instance) getRemoteWriteTimestamp() int64 { 	return ts * 1000 } +// Config returns the current Config stored by the Instance.+func (i *Instance) Config() Config {+	i.cfgMutex.Lock()+	defer i.cfgMutex.Unlock()+	return i.cfg

Just a naive question. Why is the lock required here? Isn't cfg a reference?

rfratto

comment created time in 2 months

PR opened boxdot/gurk-rs

Load data from embedded database.

Summary: This patch introduces sled's embedded database as a storage option for gurk.

+206 -21

0 comment

5 changed files

pr created time in 2 months

push eventjeschkies/gurk-rs

Karsten Jeschkies

commit sha dbe619aad8b380b69c3ab2c4291960f245efacdd

Load data from embedded database. Summary: This patch introduces sled's embedded database as a storage option for gurk.

view details

push time in 2 months

create barnchjeschkies/gurk-rs

branch : karsten/db

created branch time in 2 months

fork jeschkies/gurk-rs

Signal Messenger client for terminal

fork in 2 months

PR opened dcos/metronome

Reviewers
Job with dependency follow up.

Summary: This change allows parallel execution of dependent jobs but does not trigger a job twice if the parents did not run again.

Some small failure recovery is also introduced that loads the last active job run of a dependent job.

+88 -60

0 comment

2 changed files

pr created time in 2 months

create barnchdcos/metronome

branch : karsten/dependencies-multi-job

created branch time in 2 months

delete branch dcos/metronome

delete branch : karsten/execute-dependent-job

delete time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha b336ea0c4b4b6720ec6a4ac750985049754e92a2

Execute dependent job when all dependencies finished. (#382) Summary: This the business logic of MIP 0008 "Jobs with Dependencies". We had an actor analog to the `JobSpecSchedulerActor`. This new actor subscribes to all job run events and updates an internal state. Once all parents have a successful run that is new than the child run we trigger new job execution. JIRA issues: DCOS_OSS-5974

view details

push time in 2 months

PR merged dcos/metronome

Reviewers
Execute dependent job when all dependencies finished.

Summary: This the business logic of MIP 0008 "Jobs with Dependencies". We had an actor analog to the JobSpecSchedulerActor. This new actor subscribes to all job run events and updates an internal state. Once all parents have a successful run that is new than the child run we trigger new job execution.

JIRA issues: DCOS_OSS-5974

+355 -7

0 comment

7 changed files

jeschkies

pr closed time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha 82dc6938f90a3e9c271b332dc589c48f678fc1c5

fixup! proper error message

view details

push time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha 936bbf5702bb8564913ff8a65184d774643dee0f

fixup! renaming

view details

push time in 2 months

Pull request review commentdcos/metronome

Execute dependent job when all dependencies finised.

+package dcos.metronome+package jobspec.impl++import java.time.Instant++import akka.actor.{Actor, ActorLogging, Props, Stash, Status}+import akka.pattern.pipe+import com.typesafe.scalalogging.StrictLogging+import dcos.metronome.jobrun.{JobRunService, StartedJobRun}+import dcos.metronome.jobspec.impl.JobSpecDependencyActor.{DependenciesState, UpdateJobSpec}+import dcos.metronome.model.{Event, JobId, JobRun, JobRunStatus, JobSpec}++import scala.collection.mutable+import scala.concurrent.ExecutionContext++/**+  * Manages one JobSpec.+  *+  * If the JobSpec has dependencies, it subscribes to the events of its dependencies, ie parents.+  *+  * This actor is analog to [[JobSpecSchedulerActor]].+  *+  * TODO: find a better name+  */+class JobSpecDependencyActor(initSpec: JobSpec, runService: JobRunService) extends Actor with Stash with ActorLogging {++  private[impl] var spec = initSpec+  private[impl] var lastSuccessfulRun: Instant = Instant.MIN

Should load the last job run as a failover pattern?

jeschkies

comment created time in 2 months

Pull request review commentdcos/metronome

Execute dependent job when all dependencies finised.

+package dcos.metronome+package jobspec.impl++import java.time.Instant++import akka.actor.{Actor, ActorLogging, Props, Stash, Status}+import akka.pattern.pipe+import com.typesafe.scalalogging.StrictLogging+import dcos.metronome.jobrun.{JobRunService, StartedJobRun}+import dcos.metronome.jobspec.impl.JobSpecDependencyActor.{DependenciesState, UpdateJobSpec}+import dcos.metronome.model.{Event, JobId, JobRun, JobRunStatus, JobSpec}++import scala.collection.mutable+import scala.concurrent.ExecutionContext++/**+  * Manages one JobSpec.+  *+  * If the JobSpec has dependencies, it subscribes to the events of its dependencies, ie parents.+  *+  * This actor is analog to [[JobSpecSchedulerActor]].+  *+  * TODO: find a better name+  */+class JobSpecDependencyActor(initSpec: JobSpec, runService: JobRunService) extends Actor with Stash with ActorLogging {++  private[impl] var spec = initSpec+  private[impl] var lastSuccessfulRun: Instant = Instant.MIN+  val dependenciesState = DependenciesState(initSpec.dependencies.toSet)++  implicit val ec: ExecutionContext = context.dispatcher++  override def preStart(): Unit = {+    super.preStart()+    context.system.eventStream.subscribe(self, classOf[Event.JobRunEvent])+  }++  override def postStop(): Unit = {+    context.system.eventStream.unsubscribe(self)+    super.postStop()+  }++  override def receive: Receive = idling++  def idling: Receive = {+    case ev: Event.JobRunEvent =>+      dependenciesState.update(ev.jobRun)+      if (dependenciesState.shouldTriggerJob(lastSuccessfulRun)) {+        log.info(s"Start next run of job ${spec.id}, all parents finished successfully")+        runService.startJobRun(initSpec).pipeTo(self)+        context.become(running)

I told @pierrebeitz already what limitation we have with this.

First, I think we need to handle the result of startJobRun. It stands to question whether we need another receive method for that.

Second and this is more important, I saw job runs triggered each time there was an update. Imagine C run five seconds depends on A and B. B is run every ten seconds but A every ten millis. This would trigger C multiple times.

A way to avoid this is to save the lastCreatedAt for C and use this as well to check whether we already scheduled a run for C or not. I will create a follow up as a proposal.

jeschkies

comment created time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha 6d3fcbdc742725caf0e7baf5898d71f45682b6fd

fixup! make trigger condition easier to understand.

view details

push time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha ae2ec786505ff91f8e94c993d5c8e3da68d46ac7

Test job with dependencies integration.

view details

push time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha de03189645ce23d0017031256a287a5b5fa24936

Test job dependency actor.

view details

push time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha d76a51cc7f4f1fc576dc43a1e5c1b9fd695f4603

fixup! compile tests

view details

push time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha 31405eeb4073b7dd20dd602e3ff6269fb5b1c899

Trigger job when all its parents are successful.

view details

push time in 2 months

PR opened dcos/metronome

Reviewers
Execute dependent job when all dependencies finised.

Summary: tbd

JIRA issues: DCOS_OSS-5974

+76 -4

0 comment

3 changed files

pr created time in 2 months

create barnchdcos/metronome

branch : karsten/execute-dependent-job

created branch time in 2 months

delete branch mesosphere/marathon

delete branch : karsten/backports/1.8/MARATHON-8758

delete time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 248803b193f0ebef645706812710f4903d3b5441

Always release lock after scale check. (#7198) (#7211) Summary: `MarathonScheuduleActor` would not release the lock on a run spec if the scale check for the lock run spec failed. A scale attempt can fail when the instance of a run spec is removed before we try to set its goal. JIRA issues: MARATHON-8758 Co-authored-by: Tim Harper <tharper@mesosphere.com>

view details

push time in 2 months

PR merged mesosphere/marathon

Reviewers
Always release lock after scale check. (#7198) backport-1.8

Summary: MarathonScheuduleActor would not release the lock on a run spec if the scale check for the lock run spec failed.

A scale attempt can fail when the instance of a run spec is removed before we try to set its goal.

JIRA issues: MARATHON-8758

Co-authored-by: Tim Harper tharper@mesosphere.com

+72 -31

0 comment

1 changed file

jeschkies

pr closed time in 2 months

delete branch mesosphere/marathon

delete branch : karsten/backport/MARATHON-8758

delete time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 94002c624192b964519b8028dd9b067638738ad7

Always release lock after scale check. (#7198) (#7210) Summary: `MarathonScheuduleActor` would not release the lock on a run spec if the scale check for the lock run spec failed. A scale attempt can fail when the instance of a run spec is removed before we try to set its goal. JIRA issues: MARATHON-8758 Co-authored-by: Tim Harper <tharper@mesosphere.com>

view details

push time in 2 months

PR merged mesosphere/marathon

Reviewers
Always release lock after scale check. (#7198) backport-1.9

Summary: MarathonScheuduleActor would not release the lock on a run spec if the scale check for the lock run spec failed.

A scale attempt can fail when the instance of a run spec is removed before we try to set its goal.

JIRA issues: MARATHON-8758

Co-authored-by: Tim Harper tharper@mesosphere.com

+72 -32

0 comment

1 changed file

jeschkies

pr closed time in 2 months

pull request commentmesosphere/marathon

Allow GROUP_BY to preemptively scale up

The error message is E: Version '1.10.0-2.0.1.debian9' for 'mesos' was not found.

timcharper

comment created time in 2 months

PR opened mesosphere/marathon

Reviewers
Always release lock after scale check. (#7198)

Summary: MarathonScheuduleActor would not release the lock on a run spec if the scale check for the lock run spec failed.

A scale attempt can fail when the instance of a run spec is removed before we try to set its goal.

JIRA issues: MARATHON-8758

Co-authored-by: Tim Harper tharper@mesosphere.com

+72 -31

0 comment

1 changed file

pr created time in 2 months

create barnchmesosphere/marathon

branch : karsten/backports/1.8/MARATHON-8758

created branch time in 2 months

PR opened mesosphere/marathon

Reviewers
Always release lock after scale check. (#7198)

Summary: MarathonScheuduleActor would not release the lock on a run spec if the scale check for the lock run spec failed.

A scale attempt can fail when the instance of a run spec is removed before we try to set its goal.

JIRA issues: MARATHON-8758

Co-authored-by: Tim Harper tharper@mesosphere.com

<<Replace this line with your Revision Title>>

Summary: For more details follow http://tbaggery.com/2008/04/19/a-note-about-git-commit-messages.html.

JIRA issues:

+72 -32

0 comment

1 changed file

pr created time in 2 months

create barnchmesosphere/marathon

branch : karsten/backport/MARATHON-8758

created branch time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 0513cddcac83bd4082d48098d489e69686695998

Always release lock after scale check. (#7198) Summary: `MarathonScheuduleActor` would not release the lock on a run spec if the scale check for the lock run spec failed. A scale attempt can fail when the instance of a run spec is removed before we try to set its goal. JIRA issues: MARATHON-8758 Co-authored-by: Tim Harper <tharper@mesosphere.com>

view details

push time in 2 months

delete branch mesosphere/marathon

delete branch : karsten/scalecheck-stuck

delete time in 2 months

PR merged mesosphere/marathon

Always release lock after scale check. needs backport

Summary: MarathonScheuduleActor would not release the lock on a run spec if the scale check for the lock run spec failed.

A scale attempt can fail when the instance of a run spec is removed before we try to set its goal.

JIRA issues: MARATHON-8758

+65 -28

2 comments

1 changed file

jeschkies

pr closed time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 93d03f91b9c24f2a42c1ec374a29b89d0d67c423

fixup! comment on unprocessed message.

view details

push time in 2 months

Pull request review commentmesosphere/marathon

Always release lock after scale check.

 class MarathonSchedulerActor private (         deploy(sender(), cmd)        case DeploymentFinished(plan) =>-        removeLocks(plan.affectedRunSpecIds)         deploymentSuccess(plan)        case DeploymentFailed(plan, reason) =>-        removeLocks(plan.affectedRunSpecIds)         deploymentFailed(plan, reason) -      case RunSpecScaled(id) => removeLock(id)--      case TasksKilled(runSpecId, _) => removeLock(runSpecId)-+      case RunSpecScaled(id) =>
      case RunSpecScaled(id) => () // The lock is released via RemoveLocks.
jeschkies

comment created time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 49170a8835f13bfdd5a39fb00e244606b4bf824b

fixup! double negation

view details

push time in 2 months

Pull request review commentmesosphere/marathon

Always release lock after scale check.

 class MarathonSchedulerActor private (     conflicts.isEmpty   } -  def removeLocks(runSpecIds: Set[AbsolutePathId]): Unit = runSpecIds.foreach(removeLock)-  def removeLock(runSpecId: AbsolutePathId): Unit = {-    if (lockedRunSpecs.contains(runSpecId)) {-      val locks = lockedRunSpecs(runSpecId) - 1-      if (locks <= 0) lockedRunSpecs -= runSpecId else lockedRunSpecs(runSpecId) -= 1-      logger.debug(s"Removed lock for run spec: id=$runSpecId locks=$locks lockedRunSpec=$lockedRunSpecs")+  def removeLocks(runSpecIds: Iterable[AbsolutePathId], lockVersion: Long): Unit =+    runSpecIds.foreach { runSpecId => removeLock(runSpecId, lockVersion) }+  def removeLock(runSpecId: AbsolutePathId, lockVersion: Long): Unit = {+    // Only remove if we are at the latest lock version. Eg let's say we have two scale events for+    // app /foo after another. The first scale check locks with version 1. The seconds will lock+    // with version 2 after the lock for v1 was released. This check will then prevent that a delayed+    // duplicated lock release for v1 will not release v2.
    // duplicated lock release for v1 will release v2.
jeschkies

comment created time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 693c05e822f08c202fc6fe9a9d0f724b41e4c9f2

fixup! just some syntax changes.

view details

push time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 5b569212cc006e4205cb5bc7f3391872899ae1e3

Remove `REASON_CONTAINER_MEMORY_REQUEST_EXCEEDED`. (#7194) Summary: `REASON_CONTAINER_MEMORY_REQUEST_EXCEEDED` was removed before the actually Mesos 1.10.0 release. Marathon was depending on a snapshot release.

view details

Karsten Jeschkies

commit sha c8c0f46b8c9b3c21cf3b45f88f381a1b7d962190

Added an article on application groups (#7195) Co-authored-by: Aleksey Dukhovniy <alex.dukhovniy@googlemail.com> Co-authored-by: Tim Harper <tharper@mesosphere.com> Co-authored-by: Matthias Eichstedt <matthias.eichstedt@mesosphere.io>

view details

Karsten Jeschkies

commit sha fb2f55fa6a940c8c5054efecdbe99d23a9dfc21f

Pay attention changing minimum Mesos version. (#7191)

view details

Karsten Jeschkies

commit sha 10e31665f3505651ffbb0b2b7d28e86b43801879

Remove blocking calls from instance tracker. (#7188) Summary: Blocking calls can cause thread starvation in actors. We want to discourage this behavior by remove the last blocking calls from the instance tracker. The latest loop results ``` status error No events matched <event status_update_event (T... 1 Received an unexpected SSE event stream Connect... 1 ```

view details

Karsten Jeschkies

commit sha 4e4a53f429884c0a494db0cd8e3db6bf66e217e5

Use Scalafmt. (#7193) Summary: This formatter is used in USI and supported by IntelliJ.

view details

Karsten Jeschkies

commit sha ef85be787ec55c4b9b6e43f3793e0ce368021365

Replace thread utils with USI commons. (#7196) Summary: This is a small consolidation effort of some thread utils share with Metronome and Cosmos.

view details

Karsten Jeschkies

commit sha abb9150d7422dbf42cb9b5ad07df725affcbfed8

Merge remote-tracking branch 'origin/master' into tharper/1.10-allow-group-by-to-preemptively-scale-up

view details

push time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 1d8c2412d5c7e738d3da70a8c35389577f151c74

fixup! properly remove lock

view details

push time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha bb1d6ed9c8c1ba42171c7f703fea826fe3815b82

fixup! document lock versions.

view details

push time in 2 months

push eventmesosphere/marathon

Karsten Jeschkies

commit sha 0860ef3e21eaff472c5865e4172cb2727d9f3bae

fixup! format code.

view details

push time in 2 months

delete branch mesosphere/marathon

delete branch : karsten/use-mesosphere-proxy

delete time in 2 months

PR closed mesosphere/marathon

Use proxy for CI builds.

Summary: This configures our CI to use our Maven proxy.

+8 -1

1 comment

4 changed files

jeschkies

pr closed time in 2 months

pull request commentmesosphere/marathon

Use proxy for CI builds.

Won't do

jeschkies

comment created time in 2 months

PR closed dcos/dcos-ansible

Run build on DC/OS Jenkins.
+31 -25

0 comment

1 changed file

jeschkies

pr closed time in 2 months

PR closed jeschkies/marathon

Run pipeline in Travis CI.
+53 -14

0 comment

3 changed files

jeschkies

pr closed time in 2 months

PR closed mesosphere/usi

Reviewers
Fuzz test scheduler end to end.

Summary: This changes introduces some simple property based tests of the scheduler.

+215 -23

28 comments

9 changed files

jeschkies

pr closed time in 2 months

pull request commentmesosphere/usi

Fuzz test scheduler end to end.

Won't do.

jeschkies

comment created time in 2 months

pull request commentdcos/shakedown

Merge back Marathon patches.

Won't do I guess

jeschkies

comment created time in 2 months

PR closed mesosphere/dcos

Studio 3 Spike
+207 -1

0 comment

8 changed files

jeschkies

pr closed time in 2 months

PR closed eclipse/jgit

Follow builder pattern for RemoteAddCommand.

This allows the following chaining

Git(...).remoteAdd().setName(...).setUri(...).call()

instead of

RemoteAddCommand cmd = Git(...).remoteAdd()
cmd.setName(...)
cmd.setUri(...)
cnd.call()
+4 -2

0 comment

1 changed file

jeschkies

pr closed time in 2 months

PR closed dcos/shakedown

Reviewers
Merge back Marathon patches.

The Marathon team invested some work to make Shakedown a simple library for pytest.

Please note, this is a breaking change. We probably should bump Shakedown to 2.0.

The change:

  • Removes the logging wrapper.
  • Introduces the eventually matcher.
  • Pulls in all DC/OS client code from the CLI 0.4.
  • Replaces the spinner with retrying or eventually.
  • Replaces http.py with clean requests calls.
  • Drops all unit tests since Marathon's system test would rest all changes implicitly.
  • Authenticates with the same environment variables the CLI uses.
  • Tries the CLI session for authentication.

JIRA issue: MARATHON-8101

Co-authored-by: Alena Varkockova varkockova.a@gmail.com Co-authored-by: Aleksey Dukhovniy adukhovniy@mesosphere.io

+5222 -1963

2 comments

58 changed files

jeschkies

pr closed time in 2 months

delete branch dcos/metronome

delete branch : karsten/validate-update

delete time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha e3d6b8fe9be72bb736e60b69a217db8bd230a485

Validate job spec updates. (#381) Summary: Job spec updates were not validated for possible dependency cycles. This patch introduces the same validation already used by job spec creation. JIRA issues: D2IQ-70391

view details

push time in 2 months

PR merged dcos/metronome

Reviewers
Validate job spec updates.

Summary: Job spec updates were not validated for possible dependency cycles. This patch introduces the same validation already used by job spec creation.

JIRA issues: D2IQ-70391

+22 -5

0 comment

4 changed files

jeschkies

pr closed time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha 7ebf6f96aa3c490ce37276e534f9ebdbf94341cb

Check for dependencies when deleting job. (#380) Summary: This will check if other jobs dependent on the job we want to delete. The transaction is replaced with validations inside the job spec service actor. JIRA issues: DCOS_OSS-5978

view details

Karsten Jeschkies

commit sha edc2de150185295fed78812f6ef92f794c5e6149

Validate job spec updates. Summary: Job spec updates were not validated for possible dependency cycles. This patch introduces the same validation already used by job spec creation. JIRA issues: D2IQ-70391

view details

push time in 2 months

delete branch dcos/metronome

delete branch : karsten/delete

delete time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha 7ebf6f96aa3c490ce37276e534f9ebdbf94341cb

Check for dependencies when deleting job. (#380) Summary: This will check if other jobs dependent on the job we want to delete. The transaction is replaced with validations inside the job spec service actor. JIRA issues: DCOS_OSS-5978

view details

push time in 2 months

PR merged dcos/metronome

Reviewers
Check for dependencies when deleting job.

Summary: This will check if other jobs dependent on the job we want to delete. The transaction is replaced with validations inside the job spec service actor.

JIRA issues: DCOS_OSS-5978

+100 -82

0 comment

8 changed files

jeschkies

pr closed time in 2 months

PR opened dcos/metronome

Reviewers
Validate job spec updates.

Summary: Job spec updates were not validated for possible dependency cycles. This patch introduces the same validation already used by job spec creation.

Depends on #380

JIRA issues: D2IQ-70391

+22 -5

0 comment

4 changed files

pr created time in 2 months

create barnchdcos/metronome

branch : karsten/validate-update

created branch time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha fa948b0562516f38709fd321104b3873ff5a7b0a

Move validation into actor.

view details

push time in 2 months

PR opened dcos/metronome

Check for dependencies when deleting job.

Summary: This will check if other jobs dependent on the job we want to delete.

JIRA issues: DCOS_OSS-5978

+27 -2

0 comment

3 changed files

pr created time in 2 months

create barnchdcos/metronome

branch : karsten/delete

created branch time in 2 months

delete branch dcos/metronome

delete branch : karsten/dependencies

delete time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha 407b4306bba039d0efe86fc02fec0a9330f53005

Define dependencies on job spec. (#379) Summary: This patch introduces the `dependencies` field to job specs. Job specs cannot define dependencies on parents. JIRA issues: DCOS_OSS-5971

view details

push time in 2 months

PR merged dcos/metronome

Reviewers
Define dependencies on job spec.

Summary: This patch introduces the dependencies field to job specs. Job specs cannot define dependencies on parents.

JIRA issues: DCOS_OSS-5971

+291 -27

0 comment

18 changed files

jeschkies

pr closed time in 2 months

push eventdcos/metronome

Karsten Jeschkies

commit sha 9dd54f55ca024da868b01a7d15603569848696c8

fixup! modify in transaction.

view details

push time in 2 months

Pull request review commentdcos/metronome

Define dependencies on job spec.

 class JobSpecServiceDelegate(config: JobSpecConfig, actorRef: ActorRef, metrics:     deleteJobSpecTimeMetric {       actorRef.ask(DeleteJobSpec(id)).mapTo[JobSpec]     }++  override def transaction(updater: Seq[JobSpec] => Option[Modification]): Future[Option[JobSpec]] =+    updateJobSpecTimeMetric {+      actorRef.ask(Transaction(updater)).mapTo[JobSpec].map(Option.apply)(ExecutionContexts.callerThread)

We don't always get a job spec back.

jeschkies

comment created time in 2 months

more