【22.03SP2】【redis】使用gazelle启动redis-server，建立多个主从连接之后，主服务器下线后，从服务器需要很长时间才能重新建连

**【环境信息】**
三端环境
openeulerversion=openEuler-22.03-LTS-SP2
compiletime=2023-06-29-19-26-48
gccversion=10.3.1-37.oe2203sp2
kernelversion=5.10.0-153.12.0.92.oe2203sp2
openjdkversion=1.8.0.372.b07-1.oe2203sp2
[root@openEuler ~]# rpm -q gazelle
gazelle-1.0.2-15.aarch64
[root@openEuler ~]# rpm -q dpdk
dpdk-21.11-50.oe2203sp2.aarch64

**【问题复现步骤】**，请描述具体的操作步骤
1.主服务端启动 LD_PRELOAD=/usr/lib64/liblstack.so GAZELLE_BIND_PROCNAME=redis-server ./redis-server --protected-mode no --save "" --appendonly yes
```
dpdk_args=["--socket-mem", "1024,0,0,0", "--huge-dir", "/mnt/hugepages-lstack", "--proc-type", "primary", "--legacy-mem", "--map-perfect"]

stack_thread_mode="run-to-wakeup"

use_ltran=0
kni_switch=0

low_power_mode=0
 
#needed mbuf count = tcp_conn_count * mbuf_count_per_conn
tcp_conn_count=1500
mbuf_count_per_conn=138

# send ring size, default is 32, max is 2048
send_ring_size = 32

# 0: when send ring full, send return
# 1: when send ring full, alloc mbuf from mempool to send data
expand_send_ring = 0
 
#protocol stack thread per loop params
#read data form protocol stack into recv_ring
read_connect_number = 4
#process rpc msg number
rpc_number = 4
#read nic pkts number
nic_read_number = 128

nic_rxqueue_size = 4096
nic_txqueue_size = 2048

#each cpu core start a protocol stack thread.
num_cpus="2"

#app worker thread bind to numa in epoll/poll.
app_bind_numa=1
#app main thread affinity set by dpdk.
main_thread_affinity=0

host_addr="124.88.111.25"
mask_addr="255.255.0.0"
gateway_addr="124.88.0.1"
devices="52:54:00:f4:56:f1"

#0: use rss rule
#1: use tcp tuple rule to specify packet to nic queue
tuple_filter=0

#tuple_filter=1, below cfg valid
num_process=1
process_numa="0,1"
process_idx=0

#tuple_filer=0, below cfg valid
listen_shadow=0

#vlan mode; only support 0~4094, 0 is disabled
nic_vlan_mode=0

```

2.从服务器1（内核态）
/redis-server --protected-mode no --save "" --appendonly no
3.从服务器2（用户态）
 LD_PRELOAD=/usr/lib64/liblstack.so GAZELLE_BIND_PROCNAME=redis-server ./redis-server --protected-mode no --save "" --appendonly no

```
dpdk_args=["--socket-mem", "1024,0,0,0", "--huge-dir", "/mnt/hugepages-lstack", "--proc-type", "primary", "--legacy-mem", "--map-perfect"]

use_ltran=0
kni_switch=0

low_power_mode=0
 
#needed mbuf count = tcp_conn_count * mbuf_count_per_conn
tcp_conn_count=1500
mbuf_count_per_conn=138

# send ring size, default is 32, max is 2048
send_ring_size = 32

# 0: when send ring full, send return
# 1: when send ring full, alloc mbuf from mempool to send data
expand_send_ring = 0
 
#protocol stack thread per loop params
#read data form protocol stack into recv_ring
read_connect_number = 4
#process rpc msg number
rpc_number = 4
#read nic pkts number
nic_read_number = 128

#each cpu core start a protocol stack thread.
num_cpus="2"

#app worker thread bind to numa in epoll/poll.
app_bind_numa=1
#app main thread affinity set by dpdk.
main_thread_affinity=0

host_addr="124.88.42.223"
mask_addr="255.255.0.0"
gateway_addr="124.88.0.1"
devices="52:54:00:aa:36:f4"

#0: use rss rule
#1: use tcp tuple rule to specify packet to nic queue
tuple_filter=0

#tuple_filter=1, below cfg valid
num_process=1
process_numa="0,1"
process_idx=0

#tuple_filer=0, below cfg valid
listen_shadow=0

```
4.建立主从连接
redis-cli -h 124.88.111.25 SLAVEOF NO ONE
redis-cli -h 124.88.104.133 SLAVEOF 124.88.111.25 6379
redis-cli -h 124.88.42.223 SLAVEOF 124.88.111.25 6379

5.将主服务器下线，待连接从ESTABLISHED变成后SYN_SENT和FIN_WAIT1后，再重新启动主服务器，查看连接状态
![输入图片说明](https://foruda.gitee.com/images/1701335631200455459/f3fc9ed7_11534493.png "屏幕截图")

**【实际结果】**，请描述出问题的结果和影响
内核态可以快速建连，用户态需要等至少十分钟才能建连，一直在SYN_SENT状态

![输入图片说明](https://foruda.gitee.com/images/1701335715655953570/8c3233c8_11534493.png "屏幕截图")

**功能验证** 
当前功能ok：
第一次建立一主二从redis服务，并退出
```
409540:M 20 Aug 2024 15:08:36.896 * Ready to accept connections
409540:M 20 Aug 2024 15:11:16.925 * Replica 192.168.1.111:6379 asks for synchronization
409540:M 20 Aug 2024 15:11:16.925 * Partial resynchronization not accepted: Replication ID mismatch (Replica asked for '593b07f002bdf2bf01c64d5af327326cb673da88', my replication IDs are '47f0be10e866                                 0d9a67d4f86d497493ef4daad0d1' and '0000000000000000000000000000000000000000')
409540:M 20 Aug 2024 15:11:16.925 * Replication backlog created, my new replication IDs are '7bd1b8326ceab35f5c0775a9eb847c7b510994b6' and '0000000000000000000000000000000000000000'
409540:M 20 Aug 2024 15:11:16.925 * Starting BGSAVE for SYNC with target: disk
409540:M 20 Aug 2024 15:11:16.928 * Background saving started by pid 410703
410703:C 20 Aug 2024 15:11:16.930 * DB saved on disk
410703:C 20 Aug 2024 15:11:16.933 * RDB: 0 MB of memory used by copy-on-write
409540:M 20 Aug 2024 15:11:16.953 * Background saving terminated with success
409540:M 20 Aug 2024 15:11:16.954 * Synchronization with replica 192.168.1.111:6379 succeeded
409540:M 20 Aug 2024 15:11:19.567 * Replica 192.168.1.112:6379 asks for synchronization
409540:M 20 Aug 2024 15:11:19.567 * Partial resynchronization not accepted: Replication ID mismatch (Replica asked for 'd4b2c818007f9b4f6e0c9eb45d499d6a30a01795', my replication IDs are '7bd1b8326cea                                 b35f5c0775a9eb847c7b510994b6' and '0000000000000000000000000000000000000000')
409540:M 20 Aug 2024 15:11:19.567 * Starting BGSAVE for SYNC with target: disk
409540:M 20 Aug 2024 15:11:19.569 * Background saving started by pid 410717
410717:C 20 Aug 2024 15:11:19.571 * DB saved on disk
410717:C 20 Aug 2024 15:11:19.574 * RDB: 0 MB of memory used by copy-on-write
409540:M 20 Aug 2024 15:11:19.664 * Background saving terminated with success
409540:M 20 Aug 2024 15:11:19.665 * Synchronization with replica 192.168.1.112:6379 succeeded
^C409540:signal-handler (1724138060) Received SIGINT scheduling shutdown...
409540:M 20 Aug 2024 15:14:20.305 # User requested shutdown...
409540:M 20 Aug 2024 15:14:20.305 # Redis is now ready to exit, bye bye...
```
等待内核从redis和用户态从redis都进入SYN_SENT和FIN_WAIT1后
![](https://foruda.gitee.com/images/1724138495055901444/b852e90d_9076719.png "屏幕截图")
![输入图片说明](https://foruda.gitee.com/images/1724138481975137707/d80f11db_9076719.png "屏幕截图")
重启主服务

```
412561:M 20 Aug 2024 15:15:55.160 # Server initialized
412561:M 20 Aug 2024 15:15:55.171 * Loading RDB produced by version 6.2.9
412561:M 20 Aug 2024 15:15:55.171 * RDB age 276 seconds
412561:M 20 Aug 2024 15:15:55.171 * RDB memory usage when created 1.88 Mb
412561:M 20 Aug 2024 15:15:55.171 # Done loading RDB, keys loaded: 0, keys expired: 0.
412561:M 20 Aug 2024 15:15:55.171 * DB loaded from disk: 0.000 seconds
412561:M 20 Aug 2024 15:15:55.171 * Ready to accept connections
412561:M 20 Aug 2024 15:15:55.505 * Replica 192.168.1.111:6379 asks for synchronization
412561:M 20 Aug 2024 15:15:55.505 * Partial resynchronization not accepted: Replication ID mismatch (Replica asked for '7bd1b8326ceab35f5c0775a9eb847c7b510994b6', my replication IDs are 'e63ff156b8ec                                 08d72c968db8b807a5824f983a4e' and '0000000000000000000000000000000000000000')
412561:M 20 Aug 2024 15:15:55.505 * Replication backlog created, my new replication IDs are 'f94d1511e63b1fb8b428200414ce5dfb43a45794' and '0000000000000000000000000000000000000000'
412561:M 20 Aug 2024 15:15:55.505 * Starting BGSAVE for SYNC with target: disk
412561:M 20 Aug 2024 15:15:55.507 * Background saving started by pid 412580
412580:C 20 Aug 2024 15:15:55.509 * DB saved on disk
412580:C 20 Aug 2024 15:15:55.513 * RDB: 0 MB of memory used by copy-on-write
412561:M 20 Aug 2024 15:15:55.573 * Background saving terminated with success
412561:M 20 Aug 2024 15:15:55.574 * Synchronization with replica 192.168.1.111:6379 succeeded
412561:M 20 Aug 2024 15:15:56.328 * Replica 192.168.1.112:6379 asks for synchronization
412561:M 20 Aug 2024 15:15:56.328 * Partial resynchronization not accepted: Replication ID mismatch (Replica asked for '7bd1b8326ceab35f5c0775a9eb847c7b510994b6', my replication IDs are 'f94d1511e63b                                 1fb8b428200414ce5dfb43a45794' and '0000000000000000000000000000000000000000')
412561:M 20 Aug 2024 15:15:56.328 * Starting BGSAVE for SYNC with target: disk
412561:M 20 Aug 2024 15:15:56.330 * Background saving started by pid 412587
412587:C 20 Aug 2024 15:15:56.333 * DB saved on disk
412587:C 20 Aug 2024 15:15:56.337 * RDB: 0 MB of memory used by copy-on-write
412561:M 20 Aug 2024 15:15:56.376 * Background saving terminated with success
412561:M 20 Aug 2024 15:15:56.377 * Synchronization with replica 192.168.1.112:6379 succeeded

```

看日志时间可以看出，内核态和用户态连接都是立即完成。

src-openEuler/gazelle

内容风险标识

评论 (4)

src-openEuler/gazelle .gitee-modal { width: 500px !important; }

内容风险标识