galvesribeiro
commented
5 years ago Oh! That is critical.
The Lock async is the best option. Get Outlook for iOShttps://aka.ms/o0ukef
From: Stephen Lautier notifications@github.com Sent: Monday, June 10, 2019 8:49:41 AM To: OrleansContrib/SignalR.Orleans Cc: Subscribed Subject: [OrleansContrib/SignalR.Orleans] Messages are being dispatched multiple times (#79)
Recently I noticed that some messages are being received multiple times on the client.
After debugging a little bit, I found out that it is actually being dispatched from the server multiple times
Messages being dispatched multiple times
Streams initializing multiple times
After some digging, I believe this commit introduced the issue 66ca0aahttps://eur04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FOrleansContrib%2FSignalR.Orleans%2Fcommit%2F66ca0aa8395dd54e05f50ee93d076aa6ddc5fb61&data=02%7C01%7C%7Cdc36f4e5b103498a703508d6ed99ba1f%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636957641835044377&sdata=S9I4LF73qSmPcD%2B9CxNxklQ3%2F5wNmMJxsTFRryhXWdg%3D&reserved=0
What we can do is either call it again on the ctor, or else we can lock it? Just seen Redis implementation they do both
Redis signalr interest points
- https://github.com/aspnet/SignalR/blob/master/src/Microsoft.AspNetCore.SignalR.StackExchangeRedis/RedisHubLifetimeManager.cs#L47https://eur04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Faspnet%2FSignalR%2Fblob%2Fmaster%2Fsrc%2FMicrosoft.AspNetCore.SignalR.StackExchangeRedis%2FRedisHubLifetimeManager.cs%23L47&data=02%7C01%7C%7Cdc36f4e5b103498a703508d6ed99ba1f%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636957641835054381&sdata=X1Ueatv7dHNSH9rMpTfquGDouxqxxv%2BttbZmdVsCUIE%3D&reserved=0
- https://github.com/aspnet/SignalR/blob/master/src/Microsoft.AspNetCore.SignalR.StackExchangeRedis/RedisHubLifetimeManager.cs#L500https://eur04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Faspnet%2FSignalR%2Fblob%2Fmaster%2Fsrc%2FMicrosoft.AspNetCore.SignalR.StackExchangeRedis%2FRedisHubLifetimeManager.cs%23L500&data=02%7C01%7C%7Cdc36f4e5b103498a703508d6ed99ba1f%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636957641835064393&sdata=ikJ4wff6r7c%2BjQQyqVOWooYojnF9sZSspMaB7BLEx74%3D&reserved=0
I think there's is good because it won't lock if its already initialized.
Another note, to be honest, it's not that easily reproducible, however, on our server it happens pretty much 100% - we have 2 clients (x2 pods) and 1 client (x2 pod) connecting 6 in total
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://eur04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FOrleansContrib%2FSignalR.Orleans%2Fissues%2F79%3Femail_source%3Dnotifications%26email_token%3DABD64OFNZWXHLBTN7CUVP5TPZY5VLA5CNFSM4HWS2DAKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4GYRQGOA&data=02%7C01%7C%7Cdc36f4e5b103498a703508d6ed99ba1f%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636957641835064393&sdata=%2F1bhdCNkdWdOro4Q3ua5LTHiJEOmO8GrnrjytawUEmo%3D&reserved=0, or mute the threadhttps://eur04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FABD64OFCSQVIXDF7SA3AC7LPZY5VLANCNFSM4HWS2DAA&data=02%7C01%7C%7Cdc36f4e5b103498a703508d6ed99ba1f%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636957641835074401&sdata=PFvEqvKA7QNp8tRYZ9IlD6WFjNwHJ9ooxAzkZ44k6Tw%3D&reserved=0.
stephenlautier
Recently I noticed that some messages are being received multiple times on the client.
After debugging a little bit, I found out that it is actually being dispatched from the server multiple times
Messages being dispatched multiple times
Streams initializing multiple times
After some digging, I believe this commit introduced the issue https://github.com/OrleansContrib/SignalR.Orleans/commit/66ca0aa8395dd54e05f50ee93d076aa6ddc5fb61
What we can do is either call it again on the ctor, or else we can lock it? Just seen Redis implementation they do both
Redis signalr interest points
I think there's is good because it won't lock if its already initialized.
Another note, to be honest, it's not that easily reproducible, however, on our server it happens pretty much 100% - we have 2 clients (x2 pods) and 1 client (x2 pod) connecting 6 in total