Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Dec 13 21:37
    Aaronontheweb commented #4085
  • Dec 13 20:28
    IgorFedchenko commented #4085
  • Dec 13 20:27
    IgorFedchenko commented #4085
  • Dec 13 15:38
    Aaronontheweb labeled #4096
  • Dec 13 15:38
    Aaronontheweb milestoned #4096
  • Dec 13 15:38
    Aaronontheweb labeled #4096
  • Dec 13 15:38
    Aaronontheweb opened #4096
  • Dec 13 10:41
    peirens-bart opened #4095
  • Dec 13 08:37
    Aaronontheweb synchronize #4071
  • Dec 13 08:13
    jiyeongj opened #4094
  • Dec 12 15:42
    Aaronontheweb synchronize #4086
  • Dec 12 15:42
    Aaronontheweb closed #4083
  • Dec 12 15:42

    Aaronontheweb on dev

    Fix #4083 - Endpoint receive bu… (compare)

  • Dec 12 15:42
    Aaronontheweb closed #4089
  • Dec 12 15:42
    Aaronontheweb labeled #4093
  • Dec 12 15:42
    Aaronontheweb labeled #4093
  • Dec 12 15:42
    Aaronontheweb labeled #4093
  • Dec 12 15:42
    Aaronontheweb opened #4093
  • Dec 12 14:20
    Aaronontheweb commented #4092
  • Dec 12 14:14
    Aaronontheweb labeled #4089
Aaron Stannard
@Aaronontheweb
yep
in theory you could
nathvi
@nathvi
If possible, can you link me to those whitepapers?
Aaron Stannard
@Aaronontheweb
that's the big list for all of the Amazon ones
which runs the gamut
actually
that looks like a bunch of marketing stuff
let me find the real deal
nathvi
@nathvi
lol
here you go
that's the original technical whitepaper
not some marketing BS for people who can't figure out if they need Dynamo or Oracle
Joshua Garnett
@joshgarnett
Aaron Stannard
@Aaronontheweb
the original Dynamo concepts I think were the idea of ring-based partitioning and using that to drive replicas too
and the usage of a gossip protocol was necessary for getting all of the topology information about which replicas were / were not available
"tunable consistency"
like, I think what was original about Dynamo wasn't any one idea
nathvi
@nathvi
Thanks!
Aaron Stannard
@Aaronontheweb
but the combination of all of them
Akka.NET's "ring-based partitioning" comes from consistent hashing routers and Cluster.Sharding's use of the same algos for shard placement
doesn't quite have the same opinionated design that Dynamo does since we're not a database - we're a multi-purpose framework
nathvi
@nathvi
Aaron Stannard
@Aaronontheweb
thanks! hope you liked it
nathvi
@nathvi
I did. Though, I'm not sure I understand this...
"I can’t understate ..." Should that be "I can't overstate"?
Aaron Stannard
@Aaronontheweb
probably
I just kind of rifle those posts out extemporaneously
nathvi
@nathvi
:)
Aaron Stannard
@Aaronontheweb
and I don't go back and revise them
which... is probably not a best practice
nathvi
@nathvi
lol,
I think open source software has had a huge part in pushing Microsoft to be more open
It's hard to compete with free
Looking to make my own company.
Onur Gumus
@OnurGumus
In a cluster, if a node is temporarily down and comes back alive after it is quarantined, it cannot rejoin to the cluster. What should I do for this case ?
Joshua Garnett
@joshgarnett
Restart the node
Onur Gumus
@OnurGumus
How can I detect that is quarantined?
Joshua Garnett
@joshgarnett
I’m pretty sure events are generated
Onur Gumus
@OnurGumus
Perhaps but can't find
hmm we have this Akka.Remote.QuarantinedEvent
@joshgarnett how do I restart a node without killing the process ?
Joshua Garnett
@joshgarnett
That’s what I mean
You have to restart the process
Of the node that was quarantined
Onur Gumus
@OnurGumus
Oh ok
Onur Gumus
@OnurGumus
@joshgarnett well you can actually just restart the system without killing the process
Joshua Garnett
@joshgarnett
Fair enough, in my experience if you were quarantined it was probably for a good reason and it’s makes sense to just do a full restart of the process.
Everyone’s use case is different though