Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jun 22 18:38
    nkosi23 commented #7120
  • Jun 22 18:34
    krasin-ga commented #7099
  • Jun 22 18:30
    nkosi23 commented #7120
  • Jun 22 18:28
    nkosi23 commented #7120
  • Jun 22 18:26
    nkosi23 commented #7120
  • Jun 22 18:25
    nkosi23 commented #7120
  • Jun 22 18:25
    nkosi23 commented #7120
  • Jun 22 18:24
    nkosi23 commented #7120
  • Jun 22 18:22
    drakeforte5 commented #7120
  • Jun 22 18:21
    drakeforte5 commented #7120
  • Jun 22 18:12
    SebastianStehle commented #7120
  • Jun 22 16:43
    nkosi23 commented #7120
  • Jun 22 15:40
    ReubenBond commented #7099
  • Jun 22 15:22
    krasin-ga commented #7110
  • Jun 22 15:02
    krasin-ga commented #7099
  • Jun 22 14:59
    drakeforte5 commented #7120
  • Jun 22 14:59
    drakeforte5 commented #7120
  • Jun 22 14:58
    drakeforte5 commented #7120
  • Jun 22 14:58
    drakeforte5 commented #7120
  • Jun 22 14:57
    drakeforte5 commented #7120
Tom Nelson
@Zeroshi
ive ran it a quite a bit locally and it seems to keep consumption down
COCPORN
@COCPORN
That's great.
I've spent time implementing different protocols for logging online, and we ended up with a pretty simple dump to Cosmos in production.
I actually think I could do something good if I took a stab at it and thought differently about it.
Tom Nelson
@Zeroshi
i bet you could
i am going to play around with service fabric a little as well
COCPORN
@COCPORN
I think we all can do magical stuff if we put our minds to it and stop drinking for a fortnight.
Why?
Tom Nelson
@Zeroshi
im not sure sure what it does exactly and i would like to figure that out
COCPORN
@COCPORN
It seems to me like service fabric is dead or dying.
Tom Nelson
@Zeroshi
i think fabric supports all of azure
COCPORN
@COCPORN
I have worked with it.
Tom Nelson
@Zeroshi
whats the point of it versus orleans?
COCPORN
@COCPORN
Yes, in that way I am being hyperbolic. But starting off a new project that isn't part of the infrastructure of Azure it seems less supported than, say, AKS at this point.
I am not going to "badmouth" Orleans here. Not that I ever would. But Service Fabric seems to want to solve another problem altogether.
The same way I like to run Orleans on a Kubernetes cluster, I think it makes sense to run Orleans on SF.
Tom Nelson
@Zeroshi
yeah, it seems like it does thinks that k8 does
and monitors while performing health checks
COCPORN
@COCPORN
I would spend 5-19 minutes with the SF actor model before dismissing it.
Tom Nelson
@Zeroshi
lol
COCPORN
@COCPORN
I wrote this for fun back in the days: https://github.com/COCPORN/SimpleFabric
As a stepping stone if you wanted to create a service based on virtual actors that would easily transition to SF.
Turns out Orleans starts up in a second and it's the same code you're developing with that goes into production.
Tom Nelson
@Zeroshi
thats really cool. thank you for sharing
sorry i had a meeting
COCPORN
@COCPORN
No worries. I just start yapping some times.
Tom Nelson
@Zeroshi
why did you default the actor to 30 min? just random?
also, why isnt there an orleans' book?
COCPORN
@COCPORN
I don't understand what you mean by 30 minutes. The repo I shared is meant for exploratory coding at best, so if that is what you're referring to it is probably random.
Tom Nelson
@Zeroshi
one thing i really need to get up to speed on is the the release pipeline for orleans
COCPORN
@COCPORN
Let's do the meetup then!
I will be the heckler, you can be the voice of reason.
Tom Nelson
@Zeroshi
yeahm you just had it hard coded. i wasnt sure if your actors generally cover 90% of the needs in 30 min
i already have it on my calendar!
i ported it to linkedin as well
COCPORN
@COCPORN
In related news, I still haven't gotten a good (in my opinion) response to my idea of only releasing grains on configurable memory pressure.
Tom Nelson
@Zeroshi
define "memory pressure"
COCPORN
@COCPORN
As in: I have a bunch of grains, and they default to being retired after 2 hours of idling.
But also: I have 14GB of RAM on this computer. And retiring them when I have 10GB available is not sensible.
Tom Nelson
@Zeroshi
ah, got it
COCPORN
@COCPORN
So you can set a limit for when you start retiring grains to, say, 6GB.
Everything else will work the same.
It is literally just an if-statement in the codebase.
Tom Nelson
@Zeroshi
yeah, i like that. use a sorted list based on max idle time
COCPORN
@COCPORN
No, you don't even need to do that. It already buckets grains, so it knows which ones to retire.
Tom Nelson
@Zeroshi
i had to do that for sql cache
i meant the silo's list
COCPORN
@COCPORN
Just don't even start that process if you have a lot of memory available. This is my suggestion.
Tom Nelson
@Zeroshi
good idea
i was thinking of it as a trigger based on memory usage