redis.exceptions.ResponseError: Writing binlog failed, maybe no space left on device （主备同步无法恢复）

[fdl66]

Which PikiwiDB functionalities are relevant/related to the feature request?

No response

Description

版本： 4.0.2-alpha
#2520
此链接中提到的恢复方案——执行diskrecovery 命令， 只能恢复主机的访问， 备机同步还是异常的，无法恢复。

Proposed solution

除了清理数据后全量同步， 是否还有其他方案恢复？

Alternatives considered

有没有办法恢复主备增量同步。重新全量同步有点费劲。

[fdl66]

主机日志：

I20250213 19:32:39.849817 3056603 pika_rm.cc:69] Add Slave Node, db: (db0), ip_port: x.x.x.x:9222
I20250213 19:32:39.849833 3056603 pika_repl_server_conn.cc:156] DB: db0 TrySync Success, Session: 657
W20250213 19:32:39.903905 3056604 pika_rm.cc:163] (db0) Read Binlog error : IO error: File Does Not Exists
W20250213 19:32:39.903991 3056604 pika_repl_server_conn.cc:354] Activate Binlog Sync failed db=db0_,ip_port=x.x.x.x:9222,session id=0 IO error: File Does Not Exists
I20250213 19:32:39.904038 3056601 holy_thread.cc:293] receive noti close
I20250213 19:32:39.904093 3056601 pika_repl_server_thread.cc:24] ServerThread Close Slave Conn, fd: 286, ip_port: x.x.x.x:56460

# Replication(MASTER)
role:master
ReplicationID:513692040668115baa7603f425a1afd6a0cad4e2db31007991
connected_slaves:0
is_eligible_for_master_election:true
db0:binlog_offset=365 37748804,safety_purge=write2file355
slave_repl_offset:38310772804


root@debian:/pika/pika402/log/log_db0# ls -alh
total 941M
drwxr-xr-x  2 root root  260 Feb 13 17:41 .
drwxr-xr-x 25 root root  740 Feb 12 21:06 ..
-rw-r--r--  1 root root  64K Feb 13 17:28 manifest
-rw-r--r--  1 root root 101M Feb 13 17:38 write2file356
-rw-r--r--  1 root root 101M Feb 13 17:38 write2file357
-rw-r--r--  1 root root 101M Feb 13 17:38 write2file358
-rw-r--r--  1 root root 101M Feb 13 17:39 write2file359
-rw-r--r--  1 root root 101M Feb 13 17:39 write2file360
-rw-r--r--  1 root root 101M Feb 13 17:39 write2file361
-rw-r--r--  1 root root 101M Feb 13 17:39 write2file362
-rw-r--r--  1 root root 101M Feb 13 17:39 write2file363
-rw-r--r--  1 root root 101M Feb 13 17:39 write2file364
-rw-r--r--  1 root root  40M Feb 13 17:43 write2file365

备机日志：

==> pika.INFO <==
I20250213 19:31:31.093528 854912 pika_repl_client.cc:199] Try Send Meta Sync Request to Master (x.x.x.x:9222)
I20250213 19:31:31.095773 854861 pika_server.cc:452] Mark try connect finish
I20250213 19:31:31.095821 854861 pika_repl_client_conn.cc:155] Finish to handle meta sync response
I20250213 19:31:31.236708 854861 pika_repl_client_conn.cc:231] DB: db0 TrySync Ok


I20250213 19:31:31.290748 854860 client_thread.cc:458] Get request error 3
I20250213 19:31:31.290829 854860 client_thread.cc:465] close connection 262 reason 1 1
I20250213 19:31:31.290904 854860 pika_repl_client_thread.cc:20] ReplClient Close conn, fd=262, ip_port=x.x.x.x:11222
W20250213 19:31:31.290940 854860 pika_repl_client_thread.cc:29] Master conn disconnect : x.x.x.x:11222 try reconnect

# Replication(SLAVE)
role:slave
ReplicationID:513692040668115baa7603f425a1afd6a0cad4e2db31007991
master_host:x.x.x.x
master_port:9222
master_link_status:down
repl_connect_status:
db0:no_connect
slave_priority:100
slave_read_only:1
db_repl_state:(db0:NoConnect)
is_eligible_for_master_election:false
db0:binlog_offset=365 37748729,safety_purge=write2file355
slave_repl_offset:38310772729

root@debian:/pika/pika402/log/log_db0# ls -alh
total 937M
drwxr-xr-x  2 root root  260 Feb 13 17:41 .
drwxr-xr-x 31 root root  860 Feb 11 20:31 ..
-rw-r--r--  1 root root  64K Feb 13 17:28 manifest
-rw-r--r--  1 root root 101M Feb 13 17:40 write2file356
-rw-r--r--  1 root root 101M Feb 13 17:40 write2file357
-rw-r--r--  1 root root 101M Feb 13 17:40 write2file358
-rw-r--r--  1 root root 101M Feb 13 17:40 write2file359
-rw-r--r--  1 root root 101M Feb 13 17:40 write2file360
-rw-r--r--  1 root root 101M Feb 13 17:41 write2file361
-rw-r--r--  1 root root 101M Feb 13 17:41 write2file362
-rw-r--r--  1 root root 101M Feb 13 17:41 write2file363
-rw-r--r--  1 root root 101M Feb 13 17:41 write2file364
-rw-r--r--  1 root root  36M Feb 13 17:41 write2file365

[Mixficsol]

diskrecovery 命令是读命令，需要确认一下从节点是否执行了这个命令

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

The diskrecovery command is a read command. You need to confirm whether the slave node has executed this command.