You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have searched in the issues and found no similar issues.
What would you like to be improved?
When rss.server.heartbeat.timeout such as 60 * 1000 is bigger than rss.server.heartbeat.interval such as 10 * 1000, let's assume that just one heartbeat request was very slow effected by the network delay.
The executorService startHeartBeat execute the action of sendHearBeat by rss.server.heartbeat.interval such as 10 * 1000 milliseconds.
The executorService sendHeartBeat execute the real action of hearBeat RPC request, however when the RPC request is timeout, the thread in the executorService sendHeartBeat will be blocked. since no more two thread is available, the following sending request will be in the blockingQueue.
As a result, the server only send one hearBeat request in 60 * 1000 milliseconds, that will lead coordinator delete the server, since the server heartBeat exceeding rss.coordinator.server.heartbeat.timeout.
How should we improve?
rss.server.heartbeat.timeout should be no more than rss.server.heartbeat.interval.
Eleminate rss.server.heartbeat.timeout replaced with rss.server.heartbeat.interval.
I preffer the second resolution.
Are you willing to submit PR?
Yes I am willing to submit a PR!
The text was updated successfully, but these errors were encountered:
summaryzb
changed the title
[Improvement] rss.server.heartbeat.timeout should be no more than rss.server.heartbeat.interval
[Improvement] rss.server.heartbeat.timeout should be replaced with rss.server.heartbeat.interval
Jun 5, 2023
### What changes were proposed in this pull request?
Eleminate rss.server.heartbeat.timeout replaced with rss.server.heartbeat.interval
### Why are the changes needed?
#927
### Does this PR introduce _any_ user-facing change?
Yes, `rss.server.heartbeat.timeout` will effect nothing
### How was this patch tested?
UnitTest
Code of Conduct
Search before asking
What would you like to be improved?
When
rss.server.heartbeat.timeout
such as 60 * 1000 is bigger thanrss.server.heartbeat.interval
such as 10 * 1000, let's assume that just one heartbeat request was very slow effected by the network delay.The executorService
startHeartBeat
execute the action of sendHearBeat byrss.server.heartbeat.interval
such as 10 * 1000 milliseconds.The executorService
sendHeartBeat
execute the real action of hearBeat RPC request, however when the RPC request is timeout, the thread in the executorServicesendHeartBeat
will be blocked. since no more two thread is available, the following sending request will be in the blockingQueue.As a result, the server only send one hearBeat request in 60 * 1000 milliseconds, that will lead coordinator delete the server, since the server heartBeat exceeding
rss.coordinator.server.heartbeat.timeout
.How should we improve?
rss.server.heartbeat.timeout
should be no more thanrss.server.heartbeat.interval
.rss.server.heartbeat.timeout
replaced withrss.server.heartbeat.interval
.I preffer the second resolution.
Are you willing to submit PR?
The text was updated successfully, but these errors were encountered: