Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Mar 29 11:00
    ismaelhamed synchronize #4350
  • Mar 28 08:16
    AzureGulf commented #4360
  • Mar 28 08:14
    AzureGulf commented #4360
  • Mar 28 01:29
    Aaronontheweb commented #4360
  • Mar 28 01:25
    Aaronontheweb synchronize #4361
  • Mar 28 00:48
    Aaronontheweb synchronize #4361
  • Mar 28 00:44
    Aaronontheweb opened #4361
  • Mar 27 20:32
    Aaronontheweb milestoned #4360
  • Mar 27 20:32
    Aaronontheweb commented #4360
  • Mar 27 20:32
    Aaronontheweb assigned #4360
  • Mar 27 20:32
    Aaronontheweb labeled #4360
  • Mar 27 20:32
    Aaronontheweb labeled #4360
  • Mar 27 19:18
    dependabot-preview[bot] synchronize #3986
  • Mar 27 19:18

    dependabot-preview[bot] on nuget

    Bump NUnit from 3.6.1 to 3.12.0… (compare)

  • Mar 27 19:18
    dependabot-preview[bot] edited #3986
  • Mar 27 19:18
    dependabot-preview[bot] synchronize #4337
  • Mar 27 19:18

    dependabot-preview[bot] on nuget

    Bump FSharp.Quotations.Evaluato… (compare)

  • Mar 27 19:18
    dependabot-preview[bot] edited #4337
  • Mar 27 19:17
    dependabot-preview[bot] edited #3986
  • Mar 27 19:17
    dependabot-preview[bot] edited #4337
Joshua Garnett
@joshgarnett
Are you using akka persistence?
Ronny Carlansson
@lessismore1
Yes
Joshua Garnett
@joshgarnett
Ultimately you could have each actor store some state that specifies children. When the actor comes up it can then stand up its children actors.
So you could have an event that is something like, create child, kill child
Ronny Carlansson
@lessismore1
Context manage ActorRefs to children. Do I need to manage them in Internal State as well?
Joshua Garnett
@joshgarnett
I wouldn’t use actor refs, I’d us some other identifer
Whatever the persistence_id is of your children
Ronny Carlansson
@lessismore1
ok, thanks!
Joshua Garnett
@joshgarnett
Akka cluster sharding does some of this I believe with the remember entities setting
I haven’t dug into that code to see how it is implemented
Ronny Carlansson
@lessismore1
have a look in the code to see if I find something - thanks
Phil Sandler
@philsandler_twitter
A colleague has found that overlapping App Pool resets in IIS seem to mess up a cluster. Would that be considered a bug, or is the expectation that akka.net would not be used with IIS?
Aaron Stannard
@Aaronontheweb
@philsandler_twitter it's not a bug with Akka.NET
basically the sockets that Akka.NET has to use to stay connected to the cluster
can't have two of instances of the same socket open on the same address
which is what happens during an overlapped recycle
you'll have this issue with anything that hosts inbound socket connections
and an overlapped recycle in IIS
@goodisontoffee petabridge/akka-monitoring#37 - going to let CI run that and then I'll deploy it later today
Joshua Garnett
@joshgarnett
Has anyone built any chaos monkey style systems on top of akka.net? Thinking about injecting some latency and message loss.
Aaron Stannard
@Aaronontheweb
@joshgarnett we have a tool for that built into Akka.Remote
Joshua Garnett
@joshgarnett
oh yeah?
Aaron Stannard
@Aaronontheweb
which we mostly use inside our multi-node tests
yeah... let me see if I can dig it up
Joshua Garnett
@joshgarnett
Sweet
Aaron Stannard
@Aaronontheweb
baby sitting a nuget release on the build server here
haven't done a push of this package in 18 months
byte rot
Joshua Garnett
@joshgarnett
nice :P
Aaron Stannard
@Aaronontheweb
yeah, damn, forgot NuGet has a new explicit requirement
to specify source
gotta deal with this before I can answer your question
Joshua Garnett
@joshgarnett
no rush
Aaron Stannard
@Aaronontheweb
you enable a "transport adapter"
it's a unit that sits in front of your Akka.Remote transport
and is able to intercept messages as they're written or read
we have two built-in transport adapters we use for testing
the Gremlin, also known as the "failure injector"
which simulates packet loss
and the throttler, which simulates latency
you'd be able to setup a chaos-monkey-esque scenario using the former
it's probabilistic, FYI
you set a packet loss percentage
or you can use a
blackhole setting
which essentially forces a disconnect
Joshua Garnett
@joshgarnett
Cool, any way to also do that for local akka messages?
Aaron Stannard
@Aaronontheweb
probably by writing a custom mailbox implementation
that would behave the same way