1)其掂、簡述HA cluster原理

i). HA cluster的定義

集群cluster是指使用一組計(jì)算機(jī)作為一個(gè)整體向用戶提供一組網(wǎng)絡(luò)資源。

在集群中的每個(gè)計(jì)算機(jī)系統(tǒng)稱為集群節(jié)點(diǎn)（node）道盏。集群可隨著業(yè)務(wù)的增長柑蛇，通過添加新的節(jié)點(diǎn)的方式來提升集群性能寻馏。集群的類型包括：Load Balance翘骂、High Availability壁熄、High Performance這三種，而我們通常所說的HA cluster就是High availability cluster碳竟。

集群類型: LB(lvs/nginx(http/upstream,stream/upstream))草丧、HA、HP

ii). HA cluster 的性能衡量及工作方式

HA cluster 性能公式：

HA=MTBF/(MTBF+MTTR)*100%

MTBF: 平均故障間隔時(shí)間
MTTR: 故障的平均恢復(fù)時(shí)間
其計(jì)算值的范圍為0-1莹桅，計(jì)算得到的結(jié)果越接近1昌执，說明此HA cluster 就越穩(wěn)定。

指標(biāo): 99%,...,99.999%,99.9999%

99% 意味著一年宕機(jī)時(shí)間不超過4天；99.9% 意味一年宕機(jī)時(shí)間不超過10小時(shí)懂拾；99.99% 意味一年宕機(jī)時(shí)間不超過1小時(shí)煤禽；99.999% 意味一年宕機(jī)時(shí)間不超過6分鐘。

iii). HA cluster的工作方式

主備方式

即HA cluster集群中的節(jié)點(diǎn)以主備的方式運(yùn)行委粉，主機(jī)處于工作狀態(tài)呜师，備機(jī)處于監(jiān)控準(zhǔn)備狀態(tài)；當(dāng)主機(jī)出現(xiàn)宕機(jī)狀態(tài)時(shí)贾节，備機(jī)接管主機(jī)的一切工作， 待主機(jī)恢復(fù)正常后衷畦，備機(jī)再根據(jù)事先設(shè)置的設(shè)定來決定是否把服務(wù)切換到主機(jī)上運(yùn)行栗涂。

雙主方式

即HA cluster 集群中的節(jié)點(diǎn)均已主機(jī)方式運(yùn)行，互相之間同時(shí)運(yùn)行維護(hù)各自的服務(wù)工作并相互檢測祈争。當(dāng)任意一臺(tái)主機(jī)宕機(jī)后斤程，另一臺(tái)主機(jī)會(huì)接管它的一切工作，保證服務(wù)正常運(yùn)行菩混。

iii). HA cluster的運(yùn)行原理

自動(dòng)偵測(Auto-Detect)階段 由主機(jī)上的軟件通過冗余偵測線忿墅，經(jīng)由復(fù)雜的監(jiān)聽程序。邏輯判斷沮峡，來相互偵測對(duì)方運(yùn)行的情況疚脐，所檢查的項(xiàng)目有：主機(jī)硬件(CPU和周邊)、主機(jī)網(wǎng)絡(luò)邢疙、主機(jī)操作系統(tǒng)棍弄、數(shù)據(jù)庫引擎及其它應(yīng)用程序、主機(jī)與磁盤陣列連線疟游。為確保偵測的正確性呼畸，而防止錯(cuò)誤的判斷，可設(shè)定安全偵測時(shí)間颁虐，包括偵測時(shí)間間隔蛮原，偵測次數(shù)以調(diào)整安全系數(shù)，并且由主機(jī)的冗余通信連線另绩，將所匯集的訊息記錄下來儒陨，以供維護(hù)參考。

自動(dòng)切換(Auto-Switch)階段 某一主機(jī)如果確認(rèn)對(duì)方故障板熊，則正常主機(jī)除了繼續(xù)進(jìn)行原來的任務(wù)框全，還將依據(jù)各種容錯(cuò)備援模式接管預(yù)先設(shè)定的備援作業(yè)程序，并進(jìn)行后續(xù)的程序及服務(wù)干签，此類故障切換又被稱為failover津辩。

自動(dòng)恢復(fù)(Auto-Recovery)階段 在正常主機(jī)代替故障主機(jī)工作后，故障主機(jī)可離線進(jìn)行修復(fù)工作。在故障主機(jī)修復(fù)后喘沿，透過冗余通訊線與原正常主機(jī)連線闸度，自動(dòng)切換回修復(fù)完成的主機(jī)上。整個(gè)恢復(fù)過程完成由HA相關(guān)軟件自動(dòng)完成蚜印，亦可依據(jù)預(yù)先配置莺禁，選擇恢復(fù)動(dòng)作為半自動(dòng)或不恢復(fù)。而某資源的主節(jié)點(diǎn)故障后重新修改上線后窄赋，將轉(zhuǎn)移至其它節(jié)點(diǎn)的資源重新切回的過程通常稱為failback哟冬。

2)、keepalived實(shí)現(xiàn)主從忆绰、主主架構(gòu)

測試環(huán)境:共5臺(tái)主機(jī)
RealServer1: 192.168.10.114/24
RealServer1: 192.168.10.224/24
DirectorServer1: 192.168.10.226/24 VirtualServer: 192.168.10.10/24
DirectorServer2: 192.168.10.228/24 VirtualServer: 192.168.10.10/24

keepalived的主從架構(gòu)

i). 配置RealServer端環(huán)境

[root@rs1 ~]#ntpdate ntp1.aliyun.com
31 Dec 23:50:12 ntpdate[1617]: step time server 120.25.115.20 offset 20.688191 sec
[root@rs1 ~]#systemctl stop firewalld.service
[root@rs1 ~]#systemctl disable firewalld.service
[root@rs1 ~]#getenforce
Disabled

ii). 配置nginx測試主頁 (RS1和RS2配置類似)

[root@rs1 ~]#yum install nginx -y
[root@rs1 ~]#vim /usr/share/nginx/html/index.html
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS1_Server </h1>
[root@rs1 html]#systemctl start nginx.service
[root@rs1 html]#ss -tnl
State       Recv-Q Send-Q                   Local Address:Port                                  Peer Address:Port              
LISTEN      0      128                                  *:111                                              *:*                  
LISTEN      0      128                                  *:80

iii). 配置lvs-dr模型腳本文件

[root@rs1 html]#vim RS.sh

#!/bin/bash
#
vip=192.168.10.10
mask=255.255.255.255

case $1 in
start)
    echo 1 > /proc/sys/net/ipv4/conf/all/arp_ignore
    echo 1 > /proc/sys/net/ipv4/conf/lo/arp_ignore
    echo 2 > /proc/sys/net/ipv4/conf/all/arp_announce
    echo 2 > /proc/sys/net/ipv4/conf/lo/arp_announce
    ifconfig lo:0 $vip netmask $mask broadcast $vip up
    route add -host $vip dev lo:0
    ;;
stop)
    ifconfig lo:0 down
    echo 0 > /proc/sys/net/ipv4/conf/all/arp_ignore
        echo 0 > /proc/sys/net/ipv4/conf/lo/arp_ignore
        echo 0 > /proc/sys/net/ipv4/conf/all/arp_announce
        echo 0 > /proc/sys/net/ipv4/conf/lo/arp_announce
    ;;
*)
    echo "Usage: $(basename $0) start|stop"
    exit 1
    ;;
esac


[root@rs1 html]#bash -n RS.sh
[root@rs1 html]#bash -x RS.sh start
[root@rs1 html]#scp RS.sh 192.168.10.224:/root/

iiii). 配置DirectorServer端(DR1和DR2配置類似)

[root@dr1 ~]#ntpdate ntp.aliyun.com
 1 Jan 00:35:12 ntpdate[1653]: step time server 203.107.6.88 offset 20.667238 sec
[root@dr1 ~]#systemctl stop firewalld.service
[root@dr1 ~]#systemctl disable firewalld.service
[root@dr1 ~]#getenforce
Disabled

iiiii). 配置keepalived文件

(DR2配置需要做相應(yīng)IP的調(diào)整,包括狀態(tài)類型BACKUP以及優(yōu)先級(jí))

[root@dr1 ~]#yum install ipvsadm keepalived -y
[root@dr1 ~]#vim keepalived.conf
! Configuration File for keepalived

global_defs {
   notification_email {
    root@localhost
   }
   notification_email_from keepalived@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id 192.168.10.226
   vrrp_mcast_group4 224.0.100.19
}

vrrp_instance VI_1 {
    state MASTER
    interface ens33
    virtual_router_id 1
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 571f97b2
    }
    virtual_ipaddress {
        192.168.10.10/24 dev ens33 Label ens33:0
    }
}

virtual_server 192.168.10.10 80 {
    delay_loop 6
    lb_algo rr
    lb_kind DR
    protocol TCP

    real_server 192.168.10.114 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            }
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }
    }
    real_server 192.168.10.224 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            } 
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }   
    }   

}

[root@dr1 ~]#systemctl start keepalived
[root@dr1 ~]#ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
    link/ether 00:0c:29:66:40:a6 brd ff:ff:ff:ff:ff:ff
    inet 192.168.10.226/24 brd 192.168.10.255 scope global noprefixroute dynamic ens33
       valid_lft 11937sec preferred_lft 11937sec
    inet 192.168.10.10/24 scope global secondary ens33
       valid_lft forever preferred_lft forever

DR2同樣參照上述配置進(jìn)行設(shè)置并啟動(dòng).

iv).客戶端進(jìn)行測試

[root@CentOS6 ~]#for i in {1..20}; do curl http://192.168.10.10/index.html; done
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>

關(guān)閉dr1的keepalived服務(wù),查看dr2狀態(tài)已經(jīng)發(fā)生改變

[root@dr1 ~]#systemctl stop keepalived            
[root@dr2 keepalived]#systemctl status keepalived     
● keepalived.service - LVS and VRRP High Availability Monitor
   Loaded: loaded (/usr/lib/systemd/system/keepalived.service; disabled; vendor preset: disabled)
   Active: active (running) since 二 2019-01-01 02:17:37 CST; 8s ago
  Process: 49596 ExecStart=/usr/sbin/keepalived $KEEPALIVED_OPTIONS (code=exited, status=0/SUCCESS)
 Main PID: 49597 (keepalived)
    Tasks: 3
   CGroup: /system.slice/keepalived.service
           ├─49597 /usr/sbin/keepalived -D
           ├─49598 /usr/sbin/keepalived -D
           └─49599 /usr/sbin/keepalived -D

1月 01 02:17:37 dr2 Keepalived_vrrp[49599]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]
1月 01 02:17:42 dr2 Keepalived_vrrp[49599]: VRRP_Instance(VI_1) Transition to MASTER STATE
1月 01 02:17:43 dr2 Keepalived_vrrp[49599]: VRRP_Instance(VI_1) Entering MASTER STATE
1月 01 02:17:43 dr2 Keepalived_vrrp[49599]: VRRP_Instance(VI_1) setting protocol VIPs.
1月 01 02:17:43 dr2 Keepalived_vrrp[49599]: Sending gratuitous ARP on ens33 for 192.168.10.10
1月 01 02:17:43 dr2 Keepalived_vrrp[49599]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens33 for 192.168.10.10
1月 01 02:17:43 dr2 Keepalived_vrrp[49599]: Sending gratuitous ARP on ens33 for 192.168.10.10
1月 01 02:17:43 dr2 Keepalived_vrrp[49599]: Sending gratuitous ARP on ens33 for 192.168.10.10
1月 01 02:17:43 dr2 Keepalived_vrrp[49599]: Sending gratuitous ARP on ens33 for 192.168.10.10
1月 01 02:17:43 dr2 Keepalived_vrrp[49599]: Sending gratuitous ARP on ens33 for 192.168.10.10

查看服務(wù)調(diào)度一切正常,說明keepalived主從配置生效,反之亦然

[root@CentOS6 ~]#for i in {1..20}; do curl http://192.168.10.10/index.html; done
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>

keepalived的主主架構(gòu)

在上述基礎(chǔ)主從基礎(chǔ)做對(duì)應(yīng)的調(diào)整

i). RS方面腳本做對(duì)應(yīng)調(diào)整

[root@rs1 html]#cat RS2.sh 
#!/bin/bash
#
vip=192.168.10.99
mask=255.255.255.255

case $1 in
start)
    echo 1 > /proc/sys/net/ipv4/conf/all/arp_ignore
    echo 1 > /proc/sys/net/ipv4/conf/lo/arp_ignore
    echo 2 > /proc/sys/net/ipv4/conf/all/arp_announce
    echo 2 > /proc/sys/net/ipv4/conf/lo/arp_announce
    ifconfig lo:1 $vip netmask $mask broadcast $vip up
    route add -host $vip dev lo:1
    ;;
stop)
    ifconfig lo:1 down
    echo 0 > /proc/sys/net/ipv4/conf/all/arp_ignore
        echo 0 > /proc/sys/net/ipv4/conf/lo/arp_ignore
        echo 0 > /proc/sys/net/ipv4/conf/all/arp_announce
        echo 0 > /proc/sys/net/ipv4/conf/lo/arp_announce
    ;;
*)
    echo "Usage: $(basename $0) start|stop"
    exit 1
    ;;
esac

傳輸給RS2主機(jī),并都啟用腳本

[root@rs1 html]#scp RS2.sh 192.168.10.224:/root/
[root@rs1 html]#bash -n RS2.sh 
[root@rs1 html]#bash -x RS2.sh start

ii). DR方面對(duì)conf文件添加對(duì)應(yīng)的主備參數(shù)

DR1的配置文件:

[root@dr1 ~]#vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived

global_defs {
   notification_email {
    root@localhost
   }
   notification_email_from keepalived@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id 192.168.10.226
   vrrp_mcast_group4 224.0.100.19
}

vrrp_instance VI_1 {
    state MASTER
    interface ens33
    virtual_router_id 1
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 571f97b2
    }
    virtual_ipaddress {
        192.168.10.10
    }
}

virtual_server 192.168.10.10 80 {
    delay_loop 6
    lb_algo rr
    lb_kind DR
    protocol TCP

    real_server 192.168.10.114 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            }
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }
    }
    real_server 192.168.10.224 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            } 
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }   
    }   

}

vrrp_instance VI_2 {
    state BACKUP
    interface ens33
    virtual_router_id 2
    priority 90
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 572f97b2
    }
    virtual_ipaddress {
        192.168.10.99
    }
}

virtual_server 192.168.10.99 80 {
    delay_loop 6
    lb_algo rr
    lb_kind DR
    protocol TCP

    real_server 192.168.10.114 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            }
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }
    }
    real_server 192.168.10.224 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            }
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }
    }

}
[root@dr1 ~]#ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.10.10:80 rr
  -> 192.168.10.114:80            Route   1      0          0         
  -> 192.168.10.224:80            Route   1      0          0         
TCP  192.168.10.99:80 rr
  -> 192.168.10.114:80            Route   1      0          0         
  -> 192.168.10.224:80            Route   1      0          0 

DR2的配置文件:

[root@dr2 ~]#cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived

global_defs {
   notification_email {
    root@localhost
   }
   notification_email_from keepalived@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id 192.168.10.228
   vrrp_mcast_group4 224.0.100.19
}

vrrp_instance VI_1 {
    state BACKUP
    interface ens33
    virtual_router_id 1
    priority 90
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 571f97b2
    }
    virtual_ipaddress {
        192.168.10.10
    }
}

virtual_server 192.168.10.10 80 {
    delay_loop 6
    lb_algo rr
    lb_kind DR
    protocol TCP

    real_server 192.168.10.114 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            }
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }
    }
    real_server 192.168.10.224 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            } 
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }   
    }   

}

vrrp_instance VI_2 {
    state MASTER
    interface ens33
    virtual_router_id 2
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 572f97b2
    }
    virtual_ipaddress {
        192.168.10.99
    }
}

virtual_server 192.168.10.99 80 {
    delay_loop 6
    lb_algo rr
    lb_kind DR
    protocol TCP

    real_server 192.168.10.114 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            }
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }
    }
    real_server 192.168.10.224 80 {
        weight 1
        HTTP_GET {
            url {
              path /index.html
              status_code 200
            } 
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
        }   
    }   
    
}
[root@dr2 ~]#ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.10.10:80 rr
  -> 192.168.10.114:80            Route   1      0          0         
  -> 192.168.10.224:80            Route   1      0          0         
TCP  192.168.10.99:80 rr
  -> 192.168.10.114:80            Route   1      0          0         
  -> 192.168.10.224:80            Route   1      0          0 

iii). 客戶機(jī)測試

[root@CentOS6 ~]#for i in {1..20}; do curl http://192.168.10.10/index.html; done
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
[root@CentOS6 ~]#for i in {1..20}; do curl http://192.168.10.99/index.html; done
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>
<h1> 192.168.10.224 RS2_Server </h1>
<h1> 192.168.10.114 RS1_Server </h1>

3)浩峡、簡述http協(xié)議緩存原理及常用首部講解

程序的運(yùn)行具有局部性的特征

時(shí)間局部性: 一個(gè)數(shù)據(jù)被訪問過之后,可能很快會(huì)被再次訪問到
空間局部性: 一個(gè)數(shù)據(jù)被訪問時(shí),其周邊的數(shù)據(jù)也有可能被訪問到

cache: 命中

熱區(qū): 局部性

• 時(shí)效性:
• 緩存空間耗盡: LRU,最近最少使用算法
  • 過期: 緩存清理

緩存命中率: hit/(hit+miss)

• (0,1)
• 頁面命中率: 基于頁面數(shù)量進(jìn)行衡量
• 字節(jié)命中率: 基于頁面的體積進(jìn)行衡量

緩存與否:

• 私有數(shù)據(jù): private, private cache
• 公共數(shù)據(jù): public, public or private cache

http協(xié)議緩存的原理

http協(xié)議緩存的原理

基于nginx的反代服務(wù)時(shí)，為了加速性能错敢，可以開啟nginx緩存翰灾；如果這nginx為負(fù)載均衡器時(shí)，還要承擔(dān)緩存的功能稚茅，在高并發(fā)下纸淮，會(huì)面臨帶寬瓶頸；因此在規(guī)模交大時(shí)亚享，會(huì)在反代服務(wù)器后面添加專門用于緩存的服務(wù)器咽块，來提供緩存功能。這樣讓代理功能的服務(wù)器只負(fù)責(zé)代理虹蒋，讓緩存功能的服務(wù)器只負(fù)責(zé)緩存糜芳，當(dāng)前端主機(jī)請求資源時(shí)，它所指向的上游服務(wù)器就不在是真正的服務(wù)器魄衅，而是緩存服務(wù)器峭竣，他們之間是通過http請求和http響應(yīng)報(bào)文來通信；因此晃虫，代理服務(wù)器取資源時(shí)緩存服務(wù)器如果本地未能命中皆撩，會(huì)到后端服務(wù)器讀取數(shù)據(jù)，取到數(shù)據(jù)后按照緩存策略是否可緩存哲银，如果可緩存就把數(shù)據(jù)緩存到本地扛吞，并響應(yīng)給前端主機(jī)；如果緩存服務(wù)器能命中荆责，則緩存服務(wù)器直接響應(yīng)滥比，省去了到后端讀取數(shù)據(jù)的過程

常用首部

- Cache-related Header Fields
    - The most important caching header fields are;
        - Expire: 過期時(shí)間
            - Expries: Thu, 22 Oct 2026 06:34:30 GMT
            - Cache-Control: max-age=
            - Etag
            - If-None-Match
            - Last-Modified
            - If-Modified-Since
            - Vary
            - Age
    - 緩存有效性判斷機(jī)制:
        - 過期時(shí)間: Expires
            - HTTP/1.0
                - Expires
            - HTTP/1.1
                - Cache-Control: maxage=
                - Cache-Control: s-maxage=
        - 條件式請求:
            - Last-Modified/If-Modified-Since
            - Etag/If-None-Match
        - Expires: Thu, 13 Aug 2026 02:05:12 GMT
        - Cache-Control: maxage=315360000
        - Etag:"1ec5-502264e2ae4c0"
        - Last-Modified: Wed, 03 Sep 2014 10:00:27 GMT

4)、簡述回源原理和CDN常見多級(jí)緩存

一做院、CDN回源

回源原理

i). 源站內(nèi)容有更新的時(shí)候盲泛，源站主動(dòng)把內(nèi)容推送到CDN節(jié)點(diǎn)濒持。

ii). 常規(guī)的CDN都是回源的。即：當(dāng)有用戶訪問某一個(gè)URL的時(shí)候寺滚，如果被解析到的那個(gè)CDN節(jié)點(diǎn)沒有緩存響應(yīng)的內(nèi)容柑营，或者是緩存已經(jīng)到期，就會(huì)回源站去獲取村视。如果沒有人訪問官套，那么CDN節(jié)點(diǎn)不會(huì)主動(dòng)去源站拿的。

iii). 回源域名一般是cdn領(lǐng)域的專業(yè)術(shù)語蚁孔，通常情況下奶赔，是直接用ip進(jìn)行回源的，但是如果客戶源站有多個(gè)ip勒虾，并且ip地址會(huì)經(jīng)常變化纺阔，對(duì)于cdn廠商來說，為了避免經(jīng)常更改配置（回源ip）修然，會(huì)采用回源域名方式進(jìn)行回源，這樣即使源站的ip變化了质况，也不影響原有的配置愕宋。

二、CDN常見多級(jí)緩存

1结榄、CDN概念

• CDN的全稱是Content Delivery Network中贝，即內(nèi)容分發(fā)網(wǎng)絡(luò)。其基本思路是盡可能避開互聯(lián)網(wǎng)上有可能影響數(shù)據(jù)傳輸速度和穩(wěn)定性的瓶頸和環(huán)節(jié)臼朗，使內(nèi)容傳輸?shù)母炝谑佟⒏€(wěn)定。通過在網(wǎng)絡(luò)各處放置節(jié)點(diǎn)服務(wù)器所構(gòu)成的在現(xiàn)有的互聯(lián)網(wǎng)基礎(chǔ)之上的一層智能虛擬網(wǎng)絡(luò)视哑，CDN系統(tǒng)能夠?qū)崟r(shí)地根據(jù)網(wǎng)絡(luò)流量和各節(jié)點(diǎn)的連接绣否、負(fù)載狀況以及到用戶的距離和響應(yīng)時(shí)間等綜合信息將用戶的請求重新導(dǎo)向離用戶最近的服務(wù)節(jié)點(diǎn)上。其目的是使用戶可就近取得所需內(nèi)容挡毅，解決 Internet網(wǎng)絡(luò)擁擠的狀況蒜撮，提高用戶訪問網(wǎng)站的響應(yīng)速度。

2跪呈、CDN工作方法

• 客戶端瀏覽器先檢查是否有本地緩存是否過期段磨，如果過期，則向CDN邊緣節(jié)點(diǎn)發(fā)起請求耗绿，CDN邊緣節(jié)點(diǎn)會(huì)檢測用戶請求數(shù)據(jù)的緩存是否過期苹支，如果沒有過期，則直接響應(yīng)用戶請求误阻，此時(shí)一個(gè)完成http請求結(jié)束债蜜；如果數(shù)據(jù)已經(jīng)過期晴埂，那么CDN還需要向源站發(fā)出回源請求（back to the source request）,來拉取最新的數(shù)據(jù)。CDN的典型拓?fù)鋱D如下：

CDN的典型拓?fù)鋱D

3策幼、CDN緩存

瀏覽器本地緩存失效后邑时，瀏覽器會(huì)向CDN邊緣節(jié)點(diǎn)發(fā)起請求。類似瀏覽器緩存特姐，CDN邊緣節(jié)點(diǎn)也存在著一套緩存機(jī)制晶丘。

4、CDN緩存的缺點(diǎn)

CDN的分流作用不僅減少了用戶的訪問延時(shí)唐含，也減少的源站的負(fù)載浅浮。但其缺點(diǎn)也很明顯：當(dāng)網(wǎng)站更新時(shí)，如果CDN節(jié)點(diǎn)上數(shù)據(jù)沒有及時(shí)更新捷枯，即便用戶再瀏覽器使用Ctrl +F5的方式使瀏覽器端的緩存失效滚秩，也會(huì)因?yàn)镃DN邊緣節(jié)點(diǎn)沒有同步最新數(shù)據(jù)而導(dǎo)致用戶訪問異常。

5淮捆、CDN緩存策略

• CDN邊緣節(jié)點(diǎn)緩存策略因服務(wù)商不同而不同郁油，但一般都會(huì)遵循h(huán)ttp標(biāo)準(zhǔn)協(xié)議，通過http響應(yīng)頭中的Cache-control: max-age的字段來設(shè)置CDN邊緣節(jié)點(diǎn)數(shù)據(jù)緩存時(shí)間攀痊。

• 當(dāng)客戶端向CDN節(jié)點(diǎn)請求數(shù)據(jù)時(shí)桐腌，CDN節(jié)點(diǎn)會(huì)判斷緩存數(shù)據(jù)是否過期，若緩存數(shù)據(jù)并沒有過期苟径，則直接將緩存數(shù)據(jù)返回給客戶端案站；否則，CDN節(jié)點(diǎn)就會(huì)向源站發(fā)出回源請求棘街，從源站拉取最新數(shù)據(jù)蟆盐，更新本地緩存，并將最新數(shù)據(jù)返回給客戶端遭殉。

• CDN服務(wù)商一般會(huì)提供基于文件后綴石挂、目錄多個(gè)維度來指定CDN緩存時(shí)間，為用戶提供更精細(xì)化的緩存管理恩沽。

• CDN緩存時(shí)間會(huì)對(duì)“回源率”產(chǎn)生直接的影響誊稚。若CDN緩存時(shí)間較短，CDN邊緣節(jié)點(diǎn)上的數(shù)據(jù)會(huì)經(jīng)常失效罗心，導(dǎo)致頻繁回源里伯，增加了源站的負(fù)載，同時(shí)也增大的訪問延時(shí)渤闷；若CDN緩存時(shí)間太長疾瓮，會(huì)帶來數(shù)據(jù)更新時(shí)間慢的問題。開發(fā)者需要增對(duì)特定的業(yè)務(wù)飒箭，來做特定的數(shù)據(jù)緩存時(shí)間管理狼电。

6蜒灰、CDN緩存刷新

CDN邊緣節(jié)點(diǎn)對(duì)開發(fā)者是透明的，相比于瀏覽器Ctrl+F5的強(qiáng)制刷新來使瀏覽器本地緩存失效肩碟，開發(fā)者可以通過CDN服務(wù)商提供的“刷新緩存”接口來達(dá)到清理CDN邊緣節(jié)點(diǎn)緩存的目的强窖。這樣開發(fā)者在更新數(shù)據(jù)后，可以使用“刷新緩存”功能來強(qiáng)制CDN節(jié)點(diǎn)上的數(shù)據(jù)緩存過期削祈，保證客戶端在訪問時(shí)翅溺，拉取到最新的數(shù)據(jù)。

5)髓抑、varnish實(shí)現(xiàn)緩存對(duì)象及反代后端主機(jī)

請求報(bào)文用于通知緩存服務(wù)如何使用緩存響應(yīng)請求:

cache-request-directive =
    "no-cache"
    "no-store"
    "max-age" "=" delta-seconds
    "max-stale" [ "=" delta-seconds ]
    "min-fresh" "=" delta-seconds
    "no-transform"
    "only-if-cached"
    cache-extension

響應(yīng)報(bào)文用于通知緩存服務(wù)器如何存儲(chǔ)上級(jí)服務(wù)器響應(yīng)的內(nèi)容

cache-response-directive =
    "public"
    "public" [ "=" <"> 1#field-name <">]
    "no-cache" [ "=" <"> 1#field-name <">],可緩存,但響應(yīng)給客戶端之前需要revalidation
    "no-store", 不允許存儲(chǔ)響應(yīng)內(nèi)容于緩存中
    "no-transform"
    "must-revalidate"
    "proxy-revalidate"
    "max-age" "=" delta-seconds
    "s-maxage" "=" delta-seconds
    cache-extension

開源解決方案:

• squid:
• varnish:
• varnish官方站點(diǎn): https://varnish-cache.org/
  • Community
  • Enterprise
• 程序架構(gòu):
  • Manager進(jìn)程
  • Cache進(jìn)程,包含多種類型的線程
    • accept, worker, expiry...
  • shared memory log:
    • 統(tǒng)計(jì)數(shù)據(jù): 計(jì)數(shù)器
    • 日志區(qū)域: 日志記錄
      • varnishlog, varnishncsa, varnishstat....
  • 配置接口: VCL
    • varnish Configuration Language
      • vcl complier --> c complier --> shared object
• varnish的程序環(huán)境:
  • /etc/varnish/varnish.params: 配置varnish服務(wù)進(jìn)程的工作特性,例如監(jiān)聽的地址和端口,緩存機(jī)制
  • /etc/varnish/default.vcl: 配置各Child/Cache進(jìn)程的緩存工作屬性
  • 主程序:
    • /usr/sbin/varnishd
  • CLI interface:
    • /usr/bin/varnishadm
  • Shared Memory Log交互工具:
    • /usr/bin/varnishhist
    • /usr/bin/varnishlog
    • /usr/bin/varnishncsa
    • /usr/bin/varnishstat
    • /usr/bin/varnishtop
  • 測試工具程序
    • /usr/bin/varnishtest
  • VCL配置文件重載程序
    • /usr/sbin/varnish_reload_vcl
  • Systemd Unit File:
    • /usr/lib/systemd/system/varnish.service
      • varnish服務(wù)
    • /usr/lib/systemd/system/varnishlog.service
    • /usr/lib/systemd/system/varnishncsa.service
      • 日志持久服務(wù)
• varnish的緩存存儲(chǔ)機(jī)制（Storage Types）
  • -S [name=]type[,options]
  • malloc[,size]
    • 內(nèi)存存儲(chǔ),[,size]用于定義空間大小;重啟后所有緩存項(xiàng)失效
  • file[,path[,size[,granularity]]]
    • 磁盤文件存儲(chǔ),黑盒; 重啟后所有緩存項(xiàng)失效
  • persistent,path,size
    • 文件存儲(chǔ),黑盒; 重啟后所有緩存項(xiàng)有效; 實(shí)驗(yàn)階段
• varnish程序的選項(xiàng):
  • 程序選項(xiàng): /etc/varnish/varnish.params文件
    • -a address[:port][,address[:port]],默認(rèn)為6081端口
    • -T address[:port],默認(rèn)為6082端口
    • -s [name=]type[,options],定義緩存存儲(chǔ)機(jī)制
    • -u user
    • -g group
    • -f config: VCL配置文件
    • -F: 運(yùn)行于前臺(tái)
    • ....
  • 運(yùn)行時(shí)參數(shù): /etc/varnish/varnish.params文件, DEAMON_OPTS
    • DAEMON_OPTS="-p thread_pool_min=5 -p thread_pool_max=500 -p thread_pool_timeout=300"
    • -p param=value: 設(shè)定運(yùn)行參數(shù)及其值;可重復(fù)使用多次
    • -r param[,param...]: 設(shè)定指定的參數(shù)為只讀狀態(tài)
• 重載vcl配置文件:

~]# varnish_reload_vcl

• varnishadm

-S /etc/varnish/secret -T [ADDRESS:]PORT
help [<command>]
ping [<timestamp>]
auth <response>
quit
banner
status
start
stop
vcl.load <configname> <filename>
vcl.inline <configname> <quoted_VCLstring>
vcl.use <configname>
vcl.discard <configname>
vcl.list
param.show [-i] [<param>]
param.set <param> <value>
panic.show
panic.clear
storage.list
vcl.show [-v] <configname>
backend.list [<backend_expression>]
backend.set_health <backend_expression> <state>
ban <field> <operator> <arg> [&& <field> <oper> <arg>]...
ban.list

• 配置文件相關(guān):
  • vcl.list
  • vcl.load: 裝載, 加載并編譯
  • vcl.use: 激活
  • vcl.discard: 刪除
  • vcl.show [-v] <\configname>: 查看指定的配置文件的詳細(xì)信息
• 運(yùn)行時(shí)參數(shù):
  • param.show -l: 顯示列表
  • param.show <\PARAM>
  • param.set <\PARAM> <\VALUE>
• 緩存存儲(chǔ):
  • storage.list
• 后端服務(wù)器:
  • backend.list

VCL:

• "域"專有類型的配置語言
• state engine: 狀態(tài)引擎
• VCL有多個(gè)狀態(tài)引擎,狀態(tài)之間存在相關(guān)性,但狀態(tài)引擎彼此間互相隔離; 每個(gè)狀態(tài)引擎可使用return(x)指明關(guān)聯(lián)至哪個(gè)下一級(jí)引擎;每個(gè)狀態(tài)引擎對(duì)應(yīng)于vcl文件中的一個(gè)配置段,即為subroutine
  • vcl_hash --> return(hit) --> vcl_hit
• vcl_recv的默認(rèn)配置:

sub vcl_recv {
    if (req.method == "PRI") {
    /* We do not support SPDY or HTTP/2.0 */
    return (synth(405));
    }
    if (req.method != "GET" &&
      req.method != "HEAD" &&
      req.method != "PUT" &&
      req.method != "POST" &&
      req.method != "TRACE" &&
      req.method != "OPTIONS" &&
      req.method != "DELETE") {
        /* Non-RFC2616 or CONNECT which is weird. */
        return (pipe);
    }

    if (req.method != "GET" && req.method != "HEAD") {
        /* We only deal with GET and HEAD by default */
        return (pass);
    }
    if (req.http.Authorization || req.http.Cookie) {
        /* Not cacheable by default */
        return (pass);
    }
    return (hash);
}

• Client Side:
  • vcl_recv, vcl_pass, vcl_hit, vcl_miss, vcl_pipe, vcl_purge, vcl_synth, vcl_deliver
  • vcl_recv:
    • hash: vcl_hash
    • pass: vcl_pass
    • pipe: vcl_pipe
    • synth: vcl_synth
    • purge: vcl_hash --> vcl_purge
  • vcl_hash:
    • lookup:
      • hit: vcl_hit
      • miss: vcl_miss
      • pass, hit_for_pass: vcl_pass
      • purge: vcl_purge
• Backend Side:
  • vcl_backend_fetch, vcl_backend_response, vcl_backend_error
• 兩個(gè)特殊的引擎:
  • vcl_inti: 在處理任何請求之前要執(zhí)行的vcl代碼: 主要用于初始化VMODs;
  • vcl_fini: 所有的請求都已經(jīng)結(jié)束,在vcl配置被丟棄時(shí)調(diào)用; 主要用于清理VMODs;

vcl的語法格式:

• (1) VCL files start with vcl 4.0
• (2) //,# and /* foo */ for comments
• (3) Subroutines are declared with the sub keyword; 例如sub vcl_recv{...}
• (4) No loops, state-limited variables(受限于引擎的內(nèi)建變量)
• (5) Terminating statements with a keyword for next action as argument of the return() function, i.e:return(action); 用于實(shí)現(xiàn)狀態(tài)引擎轉(zhuǎn)換
• (6) Domain-specific

The VCL Finite State Machine

• (1) Each request is processed separately
• (2) Each request is independent from others at any given time
• (3) States are related, but isolated
• (4) return(action); exits one state and instructs Varnish to proceed to the next state
• (5) Built-in VCL code is always present and appended below your own VCL

三類主要語法

sub subroutine {
    ...
}

if CONDITION {
    ...
}else{
    ...
}

return(),hash_data()

VCL Built-in Functions and Keywords

• 函數(shù):
  • regsub(str,regex,sub)
  • regsuball(str,regex,sub)
  • ban(boolean expression)
  • hash_data(input)
  • synthetic(str)
• keywords:
  • call subroutine, return(action), new, set, unset

操作符:

• ==, !=, ~, >, >=, <, <=
• 邏輯操作符: &&, ||, !
• 變量賦值: =
• 舉例: obj,hits

if(obj.hits>0) {
    set resp.http.X-Cache = "HIT via" + server.ip;
}else{
    set resp.http.X-Cache = "MISS via" + server.ip;
}

變量類型：

內(nèi)建變量：

req.*: request, 表示由客戶端發(fā)來的請求報(bào)文相關(guān);
    req.http.*
        req.http.User-Agent, req.http.Referer,...
        
bereq.*: 由varnish發(fā)往BE主機(jī)的httpd請求先關(guān)
    bereq.http.*
    
beresp.*: 由BE主機(jī)響應(yīng)給varnish的響應(yīng)報(bào)文相關(guān)
    beresp.http.*
    
resp.*: 由varnish響應(yīng)給client相關(guān)

obj.*: 存儲(chǔ)在緩存空間中的緩存兌現(xiàn)的屬性; 只讀;

常用變量:

bereq.*,req.*:
    bereq.http.HEADERS
    bereq.request: 請求方法
    bereq.url: 請求的url
    bereq.proto: 請求的協(xié)議版本
    bereq.backend: 指明要調(diào)用的后端主機(jī)

    req.http.Cookie: 客戶端的請求報(bào)文中Cookie首部的值
    req.http.User-Agent ~ "chrome"


beresp.*.resp.*:
    beresp.http.HEADERS
    beresp.status: 響應(yīng)的狀態(tài)碼
    beresp.proto: 協(xié)議版本
    beresp.backend.name: BE主機(jī)的主機(jī)名
    beresp.ttl: BE主機(jī)響應(yīng)的內(nèi)容的余下的可緩存時(shí)長

obj.*
    obj.hits: 此對(duì)象從緩存中命中的次數(shù)
    obj.ttl: 對(duì)象的ttl值

server.*
    server.ip
    server.hostname

client.*
    client.ip

用戶自定義

• set
• unset

示例1: 強(qiáng)制對(duì)某類資源的請求不檢查緩存

vcl_recv {
    if(req.url ~ "(?i)^/(login|admin)") {
        return(pass);
    }
}

示例2: 對(duì)于特定類型的資源,例如公開的圖片等,取消其私有標(biāo)識(shí),并強(qiáng)行設(shè)定其可以由varnish緩存的時(shí)長

if(beresp.http.cache-control !~ "s-maxage") {
    if(bereq.url ~ "(?i)\.(jpg|jpeg|png|gif|css|js)$") {
        unset beresp.http.Set-Cookie;
        set beresp.ttl=3600s;
    }
}

示例3:

if(req.restarts == 0) {
    if(req.http.X-Fowarded-For) {
        set.req.http.X-Forwarded-For = req.http.X-forwarded-For + "," + client.ip;
    }else {
        set.req.http.X-Forwarded-For = client.ip;
    }
}

緩存對(duì)象的修剪: purge, ban

• (1) 能執(zhí)行purge操作

sub vcl_purge {
    return(synth(200,"Purged"));
}

• (2) 何時(shí)執(zhí)行purge操作

sub vcl_recv {
    if(req.method == "PURGE") {
        return(purge);
    }
    ...
}       

• 添加此請求的訪問控制法則:

acl purgers {
    "127.0.0.1";
    "192.168.0.0"/24;
}

sub vcl_recv {
    # allow PURGE from localhost and 192.168.0...
    if (req.method == "PURGE") {
        if (!client.ip ~ purgers) {
            return (synth(405, "Purging not allowed for " + client.ip));
        }
        return (purge);
    }
}

sub vcl_purge {
    set req.method = "GET";
    return (restart);
}

Banning

• (1)varnishadm:

ban <filed> <operator> <arg>

示例:

ban req.url ~ ^/javascripts

• (2)在配置文件中定義,使用ban()函數(shù)
  示例:

  if (req.method == "BAN") {
        ban("req.http.host == " + req.http.host + " && req.url == " + req.url);
        # Throw a synthetic page so the request won't go to the backend.
        return(synth(200, "Ban added"));
    }

11999111-18851ad8917f7b95.png

11999111-818174b37637f74f.png

如何設(shè)定使用多個(gè)后端主機(jī)

backend default {
    .host = "172.16.100.6";
    .port = "80";
}

backend appsrv {
    .host = "172.16.100.7";
    .port = "80";
}

sub req.recv {
    if(req.url ~ "(?i)\.php$") {
        set req.backend_hint = appsrv;
    }else {
        set req.backend_hint = default;
    }
    ...
}       

Director

• varnish module
  • 使用前需要導(dǎo)入
    • import directors;
      示例:

backend server1 {
    .host =
    .port =
}

backend server2 {
    .host =
    .port =
}

sub vcl_init {
    new GROUP_NAME = directors.round.robin();
    GROUP_NAME.add_backend(server1);
    GROUP_NAME.add_backend(server2);
}

sub vcl_recv {
    set req.backend_hint = GROUP_NAME.backend();
}

基于cookie的session sticky

sub vcl_init {
    new h = directors.hash();
    h.add_backend(one, 1);   // backend 'one' with weight '1'
    h.add_backend(two, 1);   // backend 'two' with weight '1'
}

sub vcl_recv {
    // pick a backend based on the cookie header of the client
    set req.backend_hint = h.backend(req.http.cookie);
}

BE Health Check

backend BE_NAME {
    .host = 
    .probe =
        .url = 
        .timeout = 
        .interval =
        .window =
        .threshold =
    }
}

• .probe: 定義健康狀態(tài)檢測方法
  • .url: 檢測時(shí)請求的URL,默認(rèn)為"/"
  • .request: 發(fā)出的具體請求
    • .request =
      • "GET /.healthtest.html HTTP/1.1"
      • "Host:www.magedu.com"
      • "Connection:close"
  • .windows: 基于最近的多少次檢查來判斷其健康狀態(tài)
  • .threshhold: 最近.window中定義的這么次檢查中只有.threshhold定義的次數(shù)是成功的
  • .interval: 檢查頻度
  • .timeout: 超時(shí)時(shí)長
  • .expected_response: 期望的響應(yīng)碼,默認(rèn)為200

健康狀態(tài)檢測的配置方式:

• (1) probe PB_NAME = {}

backend NAME = {
    .probe = PB_NAME;
    ...
}

• (2) backend NAME {}

backend NAME = {
    .probe = {
    ...
    }
}

示例:

probe check {
    .url = "/healthcheck.html";
    .timeout = 1s;
    .interval = 2s;
    .window = 5;
    .threshold = 4;
}

backend default {
    .host = "10.1.0.68";
    .port = "80";
    .probe = check;
}

backend appsrv {
    .host = "10.1.0.69";
    .port = "80";
    .probe = check;
}

[圖片上傳失敗...(image-1f6caf-1546287959236)]

設(shè)置后端的主機(jī)屬性

backend BE_NAME {
    ...
    .connect_timeout = 0.5S;
    .first_byte_timeout = 20S;
    .between_bytes_timeout = 5S;
    .max_connections = 50;
}

varnish的運(yùn)行時(shí)參數(shù):

• 線程模型:
  • cache-worker
  • cache-main
  • ban lurker
  • acceptor
  • epoll/kqueue
  • ...

線程相關(guān)的參數(shù):

在線程池內(nèi)部,其每一個(gè)請求由一個(gè)線程來處理,其worker線程的最大數(shù)決定了varnish的并發(fā)響應(yīng)能力

• thread_pools: Number of worker thread pools. 最好小于或等于CPU核心數(shù)量
• thread_pool_max: Maximum number of worker threads per pool. 每線程池的最大線程數(shù)
• thread_pool_min: Minimum number of worker threads per pool. 額外意義為"最大空閑線程數(shù)"
  • 最大并發(fā)連接數(shù)=thread_poos * thread_pool_max
• thread_pool_timeout: Period of time before idle threads are destroyed.
• thread_pool_add_delay: Period of time to wait for subsequent thread creation.
• thread_pool_destroy_delay: Added time to thread_pool_timeout.

Timer相關(guān)的參數(shù):

• send_timeout:
• timeout_idle:
• timeout_req
• 設(shè)置方式:
  • vcl.param
  • param.set
• 永久有效的方法:
  • varnish.params
    • DEAMON_OPTS="-p PARAM1=VALUE -p PARAM2=VALUE"

varnish日志區(qū)域

• shared memory log

  • 計(jì)數(shù)器
  • 日志信息

• varnishstat - Varnish Cache statistics

  • -1
  • -1 -f FILED_NAME
  • -l: 可用于-f選項(xiàng)指定的字段名稱列表
  • MAIN.cache_hit
  • MAIN.cache_miss

### varnishstat -1 -f MAIN.cache_hit -f MAIN.cache_miss
### varnishstat -l -f MAIN -f MEMPOOL

• varnishtop -Varnish log entry ranking
  • -1
  • -l taglist,可以同時(shí)使用多個(gè)-l選項(xiàng),也可以一個(gè)選項(xiàng)跟上多個(gè)標(biāo)簽
  • -I <[taglist:]regex>
  • -x taglist: 排除列表
  • -X <[taglist:]regex>
• varnishlog - Display Varnish logs
• varnishncsa - Display Varnish logs in Apache/NCSA combined log format