Skip to content
This repository has been archived by the owner on Feb 20, 2021. It is now read-only.

Fixes issue with master killing client connections while redis is write #517

Open
wants to merge 1 commit into
base: 3.0
Choose a base branch
from

Conversation

Aaron-2016
Copy link

@Aaron-2016 Aaron-2016 commented Nov 4, 2016

Fixes issue with master killing client connections while redis is under high write loads

reply_bytes was being wrapped to huge values,
e.g.omem=18446744073709551601 causing redis to close the connection as it
thought the client buffer limit had been exceeded.

See https://github.com/antirez/redis/blob/3.0/src/networking.c#L827

load.

reply_bytes was being wrapped to huge values,
e.g.omem=18446744073709551601 causing redis to close the connection as it
thought the client buffer limit had been exceeded.

See https://github.com/antirez/redis/blob/3.0/src/networking.c#L827
@ghost ghost added the cla-not-required label Nov 4, 2016
@Misterhex
Copy link

is this related to #546 ?

@Misterhex
Copy link

Misterhex commented May 3, 2017

facing similar issue here. our master is constantly dropping slave connections which in turn is causing our slaves to reconnect to master and re-syncing its replication.

replconf scheduled to be closed ASAP for overcoming of output buffer limits

@jdawgtech
Copy link

Hi, I am also constantly seeing this issue in our setup whenever client starts writing a lot of data in a short time.

the omem value reported is suspect (very large number)

can we have a release with a fix to this issue please?

using version 3.0.503
a Redis master located in Chicago Data centre (VM Windows 2012 R2), has two slaves (one in London DC, and other in Hong Kong DC)

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants