Redis trouble16 -- (Non critical) Master does not understand: Unrecognized REPLCONF option: capa

一、问题描述

看日志上边的信息有两个ERR:
1.(Non critical) Master does not understand REPLCONF capa: -ERR Unrecognized REPLCONF option: capa
2.Partial resynchronization not possible (no cached master)

1:S 19 Aug 2021 11:06:02.919 * MASTER <-> REPLICA sync started
1:S 19 Aug 2021 11:06:02.919 * Non blocking connect for SYNC fired the event.
1:S 19 Aug 2021 11:06:02.920 * Master replied to PING, replication can continue...
1:S 19 Aug 2021 11:06:02.920 * (Non critical) Master does not understand REPLCONF capa: -ERR Unrecognized REPLCONF option: capa
1:S 19 Aug 2021 11:06:02.920 * Partial resynchronization not possible (no cached master)
'全量复制开始'
1:S 19 Aug 2021 11:06:02.920 * Full resync from master: 0c7abb70be6cc347dd8590a2a31fd616c078d11c:195748251
1:M 19 Aug 2021 11:06:22.663 # Setting secondary replication ID to 9c9c7160a9e242dc83e338fbf255793652fcfe1b, valid up to offset: 1. New replication ID is de1761251cdf6a9fdd7e9de90a580840719e24a2
1:M 19 Aug 2021 11:06:22.664 * MASTER MODE enabled (user request from 'id=4858 addr=10.42.4.1:16769 fd=8 name= age=111 idle=0 flags=N db=0 sub=0 psub=0 multi=-1 qbuf=34 qbuf-free=32734 obl=0 oll=0 omem=0 events=r cmd=slaveof')
1:S 19 Aug 2021 11:11:10.794 * Before turning into a replica, using my master parameters to synthesize a cached master: I may be able to synchronize with the new master with just a partial transfer.
1:S 19 Aug 2021 11:11:10.794 * REPLICAOF 10.131.47.250:9600 enabled (user request from 'id=4866 addr=10.42.5.1:51630 fd=8 name= age=4 idle=0 flags=N db=0 sub=0 psub=0 multi=-1 qbuf=47 qbuf-free=32721 obl=0 oll=0 omem=0 events=r cmd=slaveof')
1:S 19 Aug 2021 11:11:11.422 * Connecting to MASTER 10.131.47.250:9600
1:S 19 Aug 2021 11:11:11.423 * MASTER <-> REPLICA sync started
1:S 19 Aug 2021 11:11:11.423 * Non blocking connect for SYNC fired the event.
1:S 19 Aug 2021 11:11:11.423 * Master replied to PING, replication can continue...
1:S 19 Aug 2021 11:11:11.423 * (Non critical) Master does not understand REPLCONF capa: -ERR Unrecognized REPLCONF option: capa
1:S 19 Aug 2021 11:11:11.423 * Trying a partial resynchronization (request de1761251cdf6a9fdd7e9de90a580840719e24a2:1).
'全量复制开始'
1:S 19 Aug 2021 11:11:11.424 * Full resync from master: 0c7abb70be6cc347dd8590a2a31fd616c078d11c:776617492

二、问题分析

在这种不停做全量复制的情况下,要检查主节点的日志,是否存在以下从节点被杀掉的情况
030:M 23 Nov 21:41:42.027 # Client id=28191723 addr=10.10.xx.138:38638 fd=735 name= age=119 idle=1 flags=S db=0 sub=0 psub=0 multi=-1 qbuf=0 qbuf-free=32768 obl=0 oll=10514 omem=266324615 events=rw cmd=psync scheduled to be closed ASAP for overcoming of output buffer limits.
2030:M 23 Nov 21:41:42.028 # Connection with slave 10.10.xx.138:6392 lost.
2030:M 23 Nov 21:42:24.950 * Slave 10.10.xx.138:6392 asks for synchronization
2030:M 23 Nov 21:42:24.950 * Unable to partial resync with slave 10.10.xx.138:6392 for lack of backlog (Slave request was: 3851662). 

被杀掉是因 omem=266324615 达到client-output-buffer-limit slave 512mb 128mb 60 上限

 三、问题总结

1.对于写入量大，且分片较大的情况，可以适当增大client-output-buffer-limit slave
2.减少大分片的使用，尽量每个节点控制在4GB以内。
3.可以对一些关键日志进行监控，帮助快速定位问题。