当前位置： 首页 > news >正文

keepalived原理及实战部署

news 2025/7/29 6:44:27

基于VRRP（Virtual Router Redundancy Protocol，虚拟路由冗余协议）的Keepalive机制是为了解决网络中静态默认网关单点故障的问题。在传统的网络架构中，如果作为默认网关的路由器出现故障，会导致整个子网失去对外通信的能力。而VRRP则通过引入主备模式来提供高可用性。

vrrp

VRRP（Virtual Router Redundancy Protocol，虚拟路由冗余协议）是一种网络层的容错协议。它的主要目的是在多个物理路由器之间提供一个虚拟路由器服务，以实现默认网关的高可用性。在传统的网络环境中，如果静态配置的默认网关出现故障，则会导致网络中的所有主机失去对外通信的能力。VRRP通过将多台路由器联合起来形成一个虚拟路由器，并选举出一台主路由器来处理所有的流量转发，而其他的备份路由器则随时待命，一旦主路由器出现故障，备份路由器能够迅速接替工作，从而确保网络连通性的持续性和可靠性。

工作原理

(1)虚拟路由器中的路由器根据优先级（priority）选举出Master。Master路由器通过发送免费ARP报文，将自己的虚拟MAC地址通知给与它连接的设备或者主机，从而承担报文转发任务；

(2) Master路由器周期性发送VRRP报文（心跳线），以公布其配置信息(优先级等)和工作状况；

(3) 如果Master路由器出现故障，虚拟路由器中的Backup路由器将根据优先级重新选举新的Master；

(4) 虚拟路由器状态切换时，Master路由器由一台设备切换为另外一台设备，新的Master路由器只是简单地发送一个携带虚拟路由器的MAC地址和虚拟IP地址信息的免费ARP报文，这样就可以更新与它连接的主机或设备中的ARP相关信息。网络中的主机感知不到Master路由器已经切换为另外一台设备。

(5) Backup路由器的优先级高于Master路由器时，由Backup路由器的工作方式(抢占方式和非抢占方式以及延时抢占)决定是否重新选举Master。

相关术语

虚拟路由器：Virtual Router 不是真实存在 ，虚构出来的

虚拟路由器标识：VRID(0-255)，唯一标识虚拟路由器

VIP：Virtual IP

VMAC：Virutal MAC (00-00-5e-00-01-VRID)

物理路由器：

master：主设备

backup：备用设备

priority：优先级

keepalive简介

实现原理

1.首先根据优先级选出Master提供服务

2.Master默认每间隔1秒向Baskup发送一次VRRP报文，告知Baskup自己处于工作状态，并公布配置信息(优先级等)情况

3.如果Master路由器出现故障，Backup将根据优先级重新选举为新的Master

4.当原Master恢复工作时，将根据抢占模式（抢占方式和非抢占方式以及延时抢占）决定是否重新成为Master

安装

dnf install keepalived -y

开机自启

systemctl enable --now keepalived.service

相关文件

软件包名：keepalived

主程序文件：/usr/sbin/keepalived

主配置文件：/etc/keepalived/keepalived.conf

配置文件示例：/usr/share/doc/keepalived/

Unit File：/lib/systemd/system/keepalived.service

Unit File的环境配置文件：/etc/sysconfig/keepalived

主配置文件解析

全局配置

! Configuration File for keepalivedglobal_defs {
#keepalived 发生故障切换时邮件发送的目标邮箱，可以按行区分写多个notification_email {13279419836@163.com   }#发邮件的地址notification_email_from k1@k1.org	#邮件服务器地址smtp_server 127.0.0.1#设置SMTP服务器连接超时时间，单位为秒，这里为30秒。smtp_connect_timeout 30#每个keepalived主机唯一标识，建议使用当前主机名，但多节点重名不影响router_id k1#决定了设备是否跳过对 VRRP 通告报文中源 IP 地址的检查，启用它能让设备忽略通告报文源 IP 与虚拟 IP 是否一致的验证vrrp_skip_check_adv_addr#启用严格模式，要求配置必须符合RFC文档的标准规范，否则Keepalived服务可能不会启动。#启用此项后以下状况将无法启动服务:#1.无VIP地址#2.配置了单播邻居#3.在VRRP版本2中有IPv6地址#建议关闭vrrp_strict#免费 ARP（Gratuitous ARP）报文时间间隔
#免费 ARP用于通知网络中其他设备，某 IP地址对应的 MAC 地址发生了变化
#帮助网络设备更新 ARP 缓存，确保数据能正确转发到新的主节点vrrp_garp_interval 1#用于配置发送 Gratuitous NA（免费邻居通告）报文的时间间隔  ipv6vrrp_gna_interval 1#指定组播IP地址vrrp_mcast_group4 224.0.0.44#开启keepalived执行脚本功能enable_script_security#指定脚本执行用户身份script_user root
}
#导入子配置文件
include /etc/keepalived/conf.d/*.conf

vrrp实例配置

vrrp_instance VI_1 {		#定义VRRP实例名称state MASTER			#身份为master，优先处理请求interface eth0			#指定VRRP通告报文通过哪个网络接口发送，可以和vip不在一个网卡virtual_router_id 51	#虚拟路由器惟一标识,范围：0-255，每个虚拟路由器此值必须唯一，主备需要在同一个标识符内priority 100			#当前物理节点在此虚拟路由器的优先级，范围：1-254，值越大优先级越高advert_int 1			#vrrp通告的时间间隔，默认1sauthentication {auth_type PASS|AH    #AH为IPSEC认证(不推荐),PASS为简单密码(建议使用)auth_pass 1111		#预共享密钥，仅前8位有效，保证各节点一致}track_script{			CHECK_HAPROXY		#使用脚本
}virtual_ipaddress {<IPADDR>/<MASK> brd <IPADDR> dev <STRING> scope <SCOPE> label <LABEL>#IP /掩码  		广播地址	网卡			作用域			设置别名192.168.60.100/24 dev eth0 label eth0:0  }accept 		#开启vip对外响应ping包，需要关闭vrrp_strict#默认使用nftab策略禁用ping包响应，nft list ruleset 显示策略中即可看到
}

实现单主架构

master–k1

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
global_defs {notification_email {13279419836@163.com}notification_email_from k1@k1.orgsmtp_server 127.0.0.1smtp_connect_timeout 30router_id k1vrrp_skip_check_adv_addr
#   vrrp_strictvrrp_garp_interval 1vrrp_gna_interval 1vrrp_mcast_group4 224.0.0.44
}vrrp_instance VI_1 {state MASTERinterface eth0virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1111 #主备设备认证密码需一致}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0	#vip }
}

backup–k2

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
! Configuration File for keepalivedglobal_defs {notification_email {acassen@firewall.locfailover@firewall.locsysadmin@firewall.loc}notification_email_from Alexandre.Cassen@firewall.locsmtp_server 127.0.0.1smtp_connect_timeout 30router_id k2vrrp_skip_check_adv_addr#vrrp_strictvrrp_garp_interval 1vrrp_gna_interval 1vrrp_mcast_group4 224.0.0.44
}vrrp_instance VI_1 {state BACKUP  #备interface eth0virtual_router_id 51priority 80  #优先级低于主advert_int 1authentication {auth_type PASSauth_pass 1111	#主备设备认证密码需一致}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}
}

实现效果

master设备
#查看ip归属
[root@k1 ~]# ip a
...
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000link/ether 00:0c:29:89:c7:ed brd ff:ff:ff:ff:ff:ffaltname enp3s0altname ens160inet 192.168.60.50/24 brd 192.168.60.255 scope global noprefixroute eth0valid_lft forever preferred_lft foreverinet 192.168.60.100/24 scope global secondary eth0:0 #可以看到vip此时在k1-master设备上生效valid_lft forever preferred_lft foreverinet6 fe80::869f:ac0c:2ae7:302e/64 scope link noprefixroute valid_lft forever preferred_lft forever[root@k2 ~]# ip a
...inet 192.168.60.60/24 brd 192.168.60.255 scope global noprefixroute eth0 #仅一个IP
...
#关闭k1设备keepalived
[root@k1 ~]# systemctl stop keepalived.service 
[root@k1 ~]# ip a
...inet 192.168.60.50/24 brd 192.168.60.255 scope global noprefixroute eth0   
...
[root@k2 ~]# ip a
...inet 192.168.60.60/24 brd 192.168.60.255 scope global noprefixroute eth0valid_lft forever preferred_lft foreverinet 192.168.60.100/24 scope global secondary eth0:0  	#可以看到此时vip飘到k2上valid_lft forever preferred_lft forever

非抢占模式与抢占延迟

默认为抢占模式preempt，即当高优先级的主机恢复在线后，会抢占低先级的主机的master角色，这样会使vip在KA主机中来回漂移，造成网络抖动，

建议设置为非抢占模式 nopreempt ，即高优先级主机恢复后，并不会抢占低优先级主机的master角色

非抢占模块下,如果原主机down机, VIP迁移至的新主机, 后续也发生down时,仍会将VIP迁移回原主机

注意：要关闭 VIP抢占，必须将各 keepalived 服务器state配置为BACKUP

非抢占模式

k1

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
global_defs{...}
vrrp_instance VI_1 {state BACKUP	#state设置为BACKUPinterface eth0virtual_router_id 51priority 100nopreempt		#设置为非抢占模式advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}
}

k2

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
global_defs{...}
vrrp_instance VI_1 {state BACKUPinterface eth0virtual_router_id 51priority 80nopreempt		#非抢占模式advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}
}

测试

#关闭服务
[root@k1 ~]# systemctl stop keepalived.service 
#可以看到ip飘到k2上
[root@k2 ~]# ip a
...inet 192.168.60.60/24 brd 192.168.60.255 scope global noprefixroute eth0valid_lft forever preferred_lft foreverinet 192.168.60.100/24 scope global secondary eth0:0valid_lft forever preferred_lft foreverinet6 fe80::71d1:24e2:443f:84de/64 scope link noprefixroute 
...
#重新开启服务
[root@k1 ~]# systemctl start keepalived.service 
#查看ip
[root@k1 ~]# ip a
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000link/ether 00:0c:29:89:c7:ed brd ff:ff:ff:ff:ff:ffaltname enp3s0altname ens160inet 192.168.60.50/24 brd 192.168.60.255 scope global noprefixroute eth0 #eth0网卡上仅一个ip

延迟抢占

k1

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
global_defs{...}vrrp_instance VI_1 {state BACKUPinterface eth0virtual_router_id 51priority 100preempt_delay 10		#设置抢占延迟为10秒advert_int 1authentication {auth_type PASSauth_pass 1111}

k2

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
global_defs{...}
vrrp_instance VI_1 {state BACKUPinterface eth0virtual_router_id 51priority 80preempt_delay 10advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}
}

vip单播

k1

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
global_defs{...}
vrrp_instance VI_1 {state MASTERinterface eth0virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}unicast_src_ip 192.168.60.50unicast_peer {192.168.60.60}
}

k2

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
global_defs{...}
vrrp_instance VI_1 {state BACKUPinterface eth0virtual_router_id 51priority 80preempt_delay 10advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}unicast_src_ip 192.168.60.60unicast_peer {192.168.60.50}
}

测试

[root@k1 ~]# tcpdump -i eth0 -nn src host 192.168.60.50 and dst 192.168.60.60
dropped privs to tcpdump
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), snapshot length 262144 bytes
20:53:52.038531 IP 192.168.60.50 > 192.168.60.60: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
20:53:53.044632 IP 192.168.60.50 > 192.168.60.60: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
20:53:54.049798 IP 192.168.60.50 > 192.168.60.60: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
20:53:55.050867 IP 192.168.60.50 > 192.168.60.60: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
20:53:56.057475 IP 192.168.60.50 > 192.168.60.60: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
20:53:57.061606 IP 192.168.60.50 > 192.168.60.60: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
^C
6 packets captured
6 packets received by filter
0 packets dropped by kernel

通知脚本配置

邮箱服务配置

详细配置参考配置qq邮箱收发功能，每天九点发邮件_qq邮箱每天定时发送邮件-CSDN博客

#下载s-nail
dnf install s-nail
#修改配置文件
vim /etc/s-nail.rc
# 最后一行添加以下内容：
set  from=lxx1xxxxxxxxxxx@163.com
set  smtp=smtp.163.com
set  smtp-auth-user=lxx1065372838@163.com
set  smtp-auth-password=HUaxxxxxxxxxxxxxxx
set  smtp-auth=login
# 注意：该文件输入完毕有的系统需要强制保存退出，输入wq！

mail脚本配置

vim /etc/haproxy/mail.sh#!/bin/bash
mail_dest='13279419836@qq.com'
mail_send()
{
mail_subj="$HOSTNAME to be $1 vip 转移"
mail_mess="`date +%F\ %T`: vrrp 转移,$HOSTNAME 变为 $1"
echo "$mail_mess" | mail -s "$mail_subj" $mail_dest
}
case $1 in
master)
mail_send master
;;
backup)
mail_send backup
;;
fault)
mail_send fault
;;
*)
exit 1
;;
esac#添加执行权限
chmod +x /etc/haproxy/mail.sh

haproxy.conf

global_defs {notification_email {13279419836@163.com}notification_email_from k1@k1.orgsmtp_server 127.0.0.1smtp_connect_timeout 30router_id k1vrrp_skip_check_adv_addr
#   vrrp_strictvrrp_garp_interval 1vrrp_gna_interval 1vrrp_mcast_group4 224.0.0.44enable_script_security	#开启脚本执行功能script_user root			#使用root身份执行脚本
}
vrrp_instance VI_1 {state MASTERinterface eth0virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}notify_master "/etc/keepalived/notify.sh master"   #调用脚本notify_backup "/etc/keepalived/notify.sh backup"notify_fault "/etc/keepalived/notify.sh fault"
}

测试

#重启keepalived服务
systemctl restart keepalived.service

外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传

双主架构配置

master/slave的单主架构，同一时间只有一个Keepalived对外提供服务，此主机繁忙，而另一台主机却
很空闲，利用率低下，可以使用master/master的双主架构，解决此问题。
master/master 的双主架构：
即将两个或以上VIP分别运行在不同的keepalived服务器，以实现服务器并行提供web访问的目的，提高
服务器资源利用率

k1

global_defs {notification_email {13279419836@163.com}notification_email_from k1@k1.orgsmtp_server 127.0.0.1smtp_connect_timeout 30router_id k1vrrp_skip_check_adv_addr
#   vrrp_strictvrrp_garp_interval 1vrrp_gna_interval 1vrrp_mcast_group4 224.0.0.44
}vrrp_instance VI_1 {state MASTERinterface eth0virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}
}vrrp_instance VI_60 {state BACKUPinterface eth0virtual_router_id 52priority 80    advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.200/24 dev eth0 label eth0:1}
}

k2

#打开配置文件
vim /etc/keepalived/keepalived.conf
#修改配置
global_defs{...}
vrrp_instance VI_1 {state BACKUP		interface eth0virtual_router_id 51priority 80preempt_delay 10advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.100/24 dev eth0 label eth0:0}
}
vrrp_instance VI_60 {state MASTER		interface eth0virtual_router_id 52priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.60.200/24 dev eth0 label eth0:1}
}

查看

k1:192.168.60.100
k2:192.168.60.200
[root@k1 keepalived]# ip a
...inet 192.168.60.50/24 brd 192.168.60.255 scope global noprefixroute eth0valid_lft forever preferred_lft foreverinet 192.168.60.100/24 scope global secondary eth0:0
...[root@k2 ~]# ip a
...inet 192.168.60.60/24 brd 192.168.60.255 scope global noprefixroute eth0valid_lft forever preferred_lft foreverinet 192.168.60.200/24 scope global secondary eth0:1valid_lft forever preferred_lft forever
...

ipvs高可用

双主分别实现httpd和mysql服务的调度

global_defs{...
}
vrrp_instance VI_1 {...
}
vrrp_instance VI_60 {...
}
#http
virtual_server 192.168.60.100 80 {delay_loop 6lb_algo rrlb_kind DRprotocol TCPreal_server 192.168.60.10 80 {weight 1TCP_CHECK {connect_port 80connect_timeout 5retry 3delay_before_retry 3}}real_server 192.168.60.20 80 {weight 1HTTP_GET {url {path /status_code 200}connect_timeout 1retry 3delay_before_retry 1}}}#mysql
virtual_server 192.168.60.200 3306 {delay_loop 6lb_algo rrlb_kind DRprotocol TCPreal_server 192.168.60.10 3306 {weight 1TCP_CHECK {connect_port 3306connect_timeout 10retry 3delay_before_retry 3}}real_server 192.168.60.20 3306 {weight 1TCP_CHECK {connect_timeout 10retry 3delay_before_retry 1connect_port 3306}}}

k2

global_defs{...
}
vrrp_instance VI_1 {...
}
vrrp_instance VI_60 {...
}
#http
virtual_server 192.168.60.100 80 {delay_loop 6lb_algo rrlb_kind DRpersistence_timeout 50protocol TCPreal_server 192.168.60.10 80 {weight 1TCP_CHECK {connect_port 80connect_timeout 5retry 3delay_before_retry 3}}real_server 192.168.60.20 80 {weight 1HTTP_GET {url {path /status_code 200}connect_timeout 1retry 3delay_before_retry 1}}}
#mysql
virtual_server 192.168.60.200 3306 {delay_loop 6lb_algo rrlb_kind DRprotocol TCPreal_server 192.168.60.10 3306 {weight 1TCP_CHECK {connect_port 3306connect_timeout 10retry 3delay_before_retry 3}}real_server 192.168.60.20 3306 {weight 1TCP_CHECK {connect_timeout 10retry 3delay_before_retry 1connect_port 3306}}
}

rs1/rs2

#安装nginx
dnf install nginx -y
#开启
systemctl enable --now nginx
#写入默认页面
echo rs11-192.168.60.10 > /usr/share/nginx/html/static/index.html
#安装mariadb
dnf install mariadb -y
#修改配置文件
vim /etc/my.cnf
server-id=1
#创建账户并授权
grant ALL on *.* to dll@'%' identified by 'dll'
#添加ip
ip a a 192.168.60.100/32 dev lo
ip a a 192.168.60.200/32 dev lo
#设置lo不对外响应
vim /etc/sysctl.conf
net.ipv4.conf.all.arp_ignore =1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.lo.arp_ignore =1
net.ipv4.conf.lo.arp_announce =2

测试

[root@ceshi ~]# curl 192.168.60.100
rs22-192.168.60.20
[root@ceshi ~]# curl 192.168.60.100
rs11-192.168.60.10
[root@ceshi ~]# curl 192.168.60.100
rs22-192.168.60.20
[root@ceshi ~]# curl 192.168.60.100
rs11-192.168.60.10
[root@ceshi ~]# curl 192.168.60.100
rs22-192.168.60.20[root@ceshi ~]# mysql -udll -pdll -h192.168.60.200 -e 'select @@server_id'
+-------------+
| @@server_id |
+-------------+
|          20 |
+-------------+
[root@ceshi ~]# mysql -udll -pdll -h192.168.60.200 -e 'select @@server_id'
+-------------+
| @@server_id |
+-------------+
|          10 |
+-------------+
[root@ceshi ~]# mysql -udll -pdll -h192.168.60.200 -e 'select @@server_id'
+-------------+
| @@server_id |
+-------------+
|          20 |
+-------------+
[root@ceshi ~]# mysql -udll -pdll -h192.168.60.200 -e 'select @@server_id'
+-------------+
| @@server_id |
+-------------+
|          10 |
+-------------+
[root@k1 /]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.60.100:80 rr-> 192.168.60.10:80             Route   1      0          2         -> 192.168.60.20:80             Route   1      0          3         
TCP  192.168.60.200:3306 rr-> 192.168.60.10:3306           Route   1      0          0         -> 192.168.60.20:3306           Route   1      0          0   [root@k2 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.60.100:80 rr persistent 50-> 192.168.60.10:80             Route   1      0          0         -> 192.168.60.20:80             Route   1      0          0         
TCP  192.168.60.200:3306 rr-> 192.168.60.10:3306           Route   1      0          2         -> 192.168.60.20:3306           Route   1      0          2      #关闭k1 keepalived服务 
[root@k1 /]# systemctl stop keepalived.service
测试 可以看到web服务仍可以访问
[root@ceshi ~]# curl 192.168.60.100
rs11-192.168.60.10
[root@ceshi ~]# curl 192.168.60.100
rs22-192.168.60.20
[root@ceshi ~]# curl 192.168.60.100
rs11-192.168.60.10
[root@ceshi ~]# curl 192.168.60.100
rs22-192.168.60.20
[root@ceshi ~]# curl 192.168.60.100
#在k2上查看
[root@k2 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.60.100:80 rr-> 192.168.60.10:80             Route   1      0          3         -> 192.168.60.20:80             Route   1      0          3         
TCP  192.168.60.200:3306 rr-> 192.168.60.10:3306           Route   1      0          0         -> 192.168.60.20:3306           Route   1      0          0
http://www.lryc.cn/news/602266.html

相关文章：

  • MBR和GPT分区的区别
  • 电商项目DevOps一体化运维实战
  • 【Datawhale夏令营】端侧Agent开发实践
  • CodeBuddy的安装教程
  • JAVA东郊到家按摩服务同款同城家政服务按摩私教茶艺师服务系统小程序+公众号+APP+H5
  • 基于BEKK-GARCH模型的参数估计、最大似然估计以及参数标准误估计的MATLAB实现
  • openlayer根据不同的状态显示不同的图层颜色
  • Fortran实现 3维反距离加权(IDW)插值算法
  • 初识 docker [下] 项目部署
  • ETH 交易流程深度技术详解
  • 二、Linux文本处理与文件操作核心命令
  • 从0开始学习R语言--Day60--EM插补法
  • git stash apply 冲突合并方法解决
  • Kafka 3.9.1的KRaft模式部署
  • linux系统----Ansible中的playbook简单应用
  • 从零开始的云计算生活——第三十七天，跬步千里，ansible之playbook
  • 【Blender小技巧】Blender使用多边形建形工具创建多边形模型，挤出面，模型创建修改编辑UV贴图
  • 【第四章:大模型（LLM)】01.神经网络中的 NLP-(2)Seq2Seq 原理及代码解析
  • 从0到500账号管理：亚矩阵云手机多开组队与虚拟定位实战指南
  • 【归并排序】排序数组（medium）
  • Rust/Tauri 优秀开源项目推荐
  • C/C++ 调用lua脚本，lua脚本调用另一个lua脚本
  • 最新的前端技术和趋势（2025）
  • Maven中的bom和父依赖
  • Nginx HTTP 反向代理负载均衡实验
  • YOLO11 改进、魔改｜低分辨率自注意力机制LRSA ,提取全局上下文建模与局部细节，提升小目标、密集小目标的检测能力
  • 免费 SSL 证书申请简明教程，让网站实现 HTTPS 访问
  • ADAS测试：如何用自动化手段提升VV效率
  • 【CDA干货】金融超市电商App经营数据分析案例
  • unbuntn 22.04 coreutils文件系统故障