Closed bsergean closed 4 years ago
If almost feel as if the unix socket code path was used for some reason (which would explain the No such file or directory ... with a tcp socket I don't expect that kind of error to pop up right ?).
I have bumped the number of replicas, and now things to be stable. I'm finally using the proxy and redis cluster now !!! I'm very happy :)
I have 6 replicas now, and the second one crashed and restarted 4 times.
redis-proxy-28-2qk2v 1/1 Running 1 4h
redis-proxy-28-h28fz 1/1 Running 4 5h
redis-proxy-28-hknjf 1/1 Running 0 3h
redis-proxy-28-mkg97 1/1 Running 0 4h
redis-proxy-28-n7jrv 1/1 Running 0 3h
redis-proxy-28-s5xmv 1/1 Running 0 4h
Here's the main log:
[2020-01-27 22:25:50.244] Redis Cluster Proxy v999.999.999 (unstable)
[2020-01-27 22:25:50.244] Commit: (00000000/0)
[2020-01-27 22:25:50.244] Cluster Address: 172.24.222.56:6379
[2020-01-27 22:25:50.244] PID: 1
[2020-01-27 22:25:50.244] OS: Linux 3.10.0-957.1.3.el7.x86_64 x86_64
[2020-01-27 22:25:50.244] Bits: 64
[2020-01-27 22:25:50.244] Log level: info
[2020-01-27 22:25:50.244] Connections pool size: 10 (respawn 2 every 50ms if below 10)
[2020-01-27 22:25:50.244] The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
[2020-01-27 22:25:50.244] Not listening to IPv6: unsupported
[2020-01-27 22:25:50.245] Listening on *:7777
[2020-01-27 22:25:50.245] Starting 8 threads...
[2020-01-27 22:25:50.245] Fetching cluster configuration...
[2020-01-27 22:25:50.280] Cluster has 10 masters and 10 replica(s)
Sounds like this got fixed by 0ef1e43c9294655ba30a6a625091d44f7d1781f6
I am trying to put the proxy in front of a bunch of traffic and I'm getting those errors:
I am using the connection_pool branch, and I have merged all commits from the unstable branch as of
Mon Jan 27 14:24:09 PST 2020
. I use the default config, no special args.