Redis Log Error writing to client: Broken pipe

[Abhishek Tiwari]

Hi All,

I am getting following error

8488:M 14 Oct 15:00:46.898 - Error writing to client: Broken pipe

8488:M 14 Oct 15:00:46.899 - Error writing to client: Broken pipe

8488:M 14 Oct 15:01:13.807 - Error writing to client: Connection reset by peer

8488:M 14 Oct 15:01:13.808 - Error writing to client: Broken pipe

8488:M 14 Oct 15:01:13.809 - Error writing to client: Broken pipe

8488:M 14 Oct 15:01:44.231 - Error writing to client: Broken pipe

8488:M 14 Oct 15:01:44.231 - Error writing to client: Broken pipe

What is the exact reason behind above error and how can i remove it .

[Josiah Carlson]

Your client is disconnecting for one reason or another. This could be caused by a broken firewall configuration, broken client, failing hardware, or a thousand other reasons. Looks like your connections reset roughly every 30 seconds. Do you have something with a 30 second timeout that kills your clients every 30 seconds?

 - Josiah

--
You received this message because you are subscribed to the Google Groups "Redis DB" group.
To unsubscribe from this group and stop receiving emails from it, send an email to redis-db+u...@googlegroups.com.
To post to this group, send email to redi...@googlegroups.com.
Visit this group at http://groups.google.com/group/redis-db.
For more options, visit https://groups.google.com/d/optout.

[Brijesh Bais]

Our setup of redis is like below :

1.  Client : phpredis  ( timeout on client side is 2 sec)

2.   Client connect to twemp proxy and forward the request to redis instances , we have 6 redis instances

we are using redis as a LRU cache.

given the redis and twemp proxy config in above comment.

but we are getting read time out some time very high count.

could you suggest , what can we tune or need to change in config or something else.


Thanks,
Brijesh

[Brijesh Bais]

Hi Josiah,

Abhishek is my colleague. We are working on the same thing we have not defined the 30 sec. timeout  on either twemp proxy or redis server.

below is my twemp proxy and redis config :

Twemp :

Production_Frontend:
  listen: 10.20.73.123:22022
  hash: fnv1a_64
  hash_tag: "{}"
  distribution: ketama
  timeout: 2000
  preconnect: true
  auto_eject_hosts: true
  server_retry_timeout: 20000
  server_failure_limit: 3
  server_connections: 12
  redis: true
  servers:
   - redis1:6379:1
   - redis2:6379:1
   - redis3:6379:1
   - redis4:6379:1
   - redis5:6379:1
   - redis6:6379:1

Redis Config :
daemonize yes
pidfile /var/run/redis_6379.pid
port 6379
tcp-backlog 2048
bind 10.20.72.2
timeout 0
tcp-keepalive 0
loglevel verbose
logfile /var/log/redis/redis.log
databases 16
stop-writes-on-bgsave-error yes
rdbcompression yes
rdbchecksum yes
dbfilename dump.rdb
dir /mnt/vol1/redis
maxclients 10000
maxmemory 28000000000
maxmemory-policy volatile-ttl

Thanks,
Brijesh Bais

On Wednesday, October 14, 2015 at 11:09:25 PM UTC+5:30, Josiah Carlson wrote:

[Josiah Carlson]

Have you checked the slowlog around the times where these connection failures are happening? Is there anything else running on the servers with Redis on them?

 - Josiah