Dunner1991's questions

在过去的几天里，我一直在对 2 个 DNS 服务器进行基准测试，我遇到了以下问题

当我使用 dnsperf 时，1 个 DNS 服务器的 qps per seconf 限制为 ~6500 qps，而另一个 DNS 服务器仅获得 ~1500 qps

Statistics:
Queries sent:         77511
  Queries completed:    77511 (100.00%)
  Queries lost:         0 (0.00%)

  Response codes:       NOERROR 69982 (90.29%), NXDOMAIN 7529 (9.71%)
  Average packet size:  request 48, response 281
  Run time (s):         11.473787
  Queries per second:   6755.485351

  Average Latency (s):  1.190392 (min 0.000293, max 1.753760)
  Latency StdDev (s):   0.443875



Statistics:

  Queries sent:         34372
  Queries completed:    14676 (42.70%)
  Queries lost:         19696 (57.30%)

  Response codes:       NOERROR 13246 (90.26%), NXDOMAIN 1430 (9.74%)
  Average packet size:  request 47, response 247
  Run time (s):         10.132536
  Queries per second:   1448.403440

  Average Latency (s):  0.111563 (min 0.000680, max 1.615017)
  Latency StdDev (s):   0.050619

两个 DNS 服务器都是相同的 centos 和绑定版本，它们都有 16 个 vCPU 和 16 GB RAM 它们位于不同的数据中心，每个我都有为每个本地构建的 dnsperf（基准服务器）。

有没有办法增加这个 qps 限制或解决它的方法？

我正在推出外部 DNS 服务器以解决所有最终用户查询

在我们将解决方案投入生产之前，我希望尽可能多地进行有用的日志记录

这是 named.conf 文件（命名空间称为 gi-named.conf 文件）

options {
        listen-on port 53 { Public IP; };
        #listen-on-v6 port 53 { ::1; };
        directory       "/var/named";
        dump-file       "/var/named/data/cache_dump.db";
        statistics-file "/var/named/data/named_stats.txt";
        memstatistics-file "/var/named/data/named_mem_stats.txt";
        recursing-file  "/var/named/data/named.recursing";
        secroots-file   "/var/named/data/named.secroots";
        allow-query     { any; };
        allow-query-on  { PublicIP; };

        /*
         - If you are building an AUTHORITATIVE DNS server, do NOT enable recursion.
         - If you are building a RECURSIVE (caching) DNS server, you need to enable
           recursion.
         - If your recursive DNS server has a public IP address, you MUST enable access
           control to limit queries to your legitimate users. Failing to do so will
           cause your server to become part of large scale DNS amplification
           attacks. Implementing BCP38 within your network would greatly
           reduce such attack surface
        */
        recursion yes;
        allow-query-cache { Internal Range; };
        allow-query-cache-on  { PublicIP; };



        query-source address Public IP ;

        dnssec-enable yes;
        dnssec-validation yes;

        /* Path to ISC DLV key */
        bindkeys-file "/etc/named.iscdlv.key";

        managed-keys-directory "/var/named/dynamic";

        pid-file "/run/named/named.pid";
        session-keyfile "/run/named/session.key";
};


logging
{
/*      If you want to enable debugging, eg. using the 'rndc trace' command,
 *      named will try to write the 'named.run' file in the $directory (/var/named).
 *      By default, SELinux policy does not allow named to modify the /var/named directory,
 *      so put the default debug log file in data/ :
 */
        /*channel default_debug {
                print-time yes;
                print-category yes;
                print-severity yes;
                file "data/named.run";
                severity dynamic;
        };*/
        channel queries_log {
                file "/var/log/queries" versions 1 size 20m;
                print-time yes;
                print-category yes;
                print-severity yes;
                severity debug 3;
        };
        channel default_log {
           file "/var/named/log/default" versions 3 size 20m;
           print-time yes;
           print-category yes;
           print-severity yes;
           severity info;
    };
       channel query-errors_log {
           file "/var/named/log/query-errors" versions 5 size 20m;
           print-time yes;
           print-category yes;
           print-severity yes;
           severity dynamic;
    };

        category queries { queries_log; };
        category client { queries_log;  };
};

zone "." IN {
        type hint;
        file "named.ca";
};

include "/etc/named.rfc1912.zones";
include "/etc/named.root.key";

我遇到的问题是，当我看到填充 /var/named/log/queries 文件的日志时，我在 /var/named/log/query-error 日志文件或 /var/named/log 中都看不到任何日志/default 日志文件，我不确定我到底哪里出错了，或者我是否遗漏了一些配置

任何人以前经历过这种情况吗？

我有一个位于公共 IP 后面的 DNS 服务器集群。

这些服务器有时会解决，但有时它们只是为任何查询返回一个 ServFail 错误代码

我的设置不是典型的（这是继承的）。

基本上在服务器上有一个名为 gi 的命名空间，这里是新服务调用 srv-gi ''' 使用命名服务的地方

#!/bin/sh

start_service() {
        ip netns exec gi /usr/sbin/zebra -d -A 127.0.0.1 -f /etc/quagga/zebra.conf
        ip netns exec gi /usr/sbin/bgpd -d -A 127.0.0.1 -f /etc/quagga/bgpd.conf 
        #DNS service
        ip netns exec gi  /usr/sbin/named -u named -c /etc/gi-named.conf
}

start_service

'''

named.conf 文件也已重命名为 gi-named.conf 文件。

// // named.conf // // 由 Red Hat 绑定包提供，用于将 ISC BIND named(8) DNS // 服务器配置为仅缓存名称服务器（仅作为 localhost DNS 解析器）。// // 参见 /usr/share/doc/bind*/sample/ 例如命名的配置文件。// // 有关位于 /usr/share/doc/bind-{version}/Bv9ARM.html 中的配置的详细信息，请参阅 BIND 管理员参考手册 (ARM)

options {
        listen-on port 53 { Public IP; };
        #listen-on-v6 port 53 { ::1; };
        directory       "/var/named";
        dump-file       "/var/named/data/cache_dump.db";
        statistics-file "/var/named/data/named_stats.txt";
        memstatistics-file "/var/named/data/named_mem_stats.txt";
        recursing-file  "/var/named/data/named.recursing";
        secroots-file   "/var/named/data/named.secroots";
        allow-query     { any; };
        allow-query-on  { PublicIP; };

        /*
         - If you are building an AUTHORITATIVE DNS server, do NOT enable recursion.
         - If you are building a RECURSIVE (caching) DNS server, you need to enable
           recursion.
         - If your recursive DNS server has a public IP address, you MUST enable access
           control to limit queries to your legitimate users. Failing to do so will
           cause your server to become part of large scale DNS amplification
           attacks. Implementing BCP38 within your network would greatly
           reduce such attack surface
        */
        recursion yes;
        allow-query-cache { Internal Range; };
        allow-query-cache-on  { PublicIP; };



        query-source address Public IP ;

        dnssec-enable yes;
        dnssec-validation yes;

        /* Path to ISC DLV key */
        bindkeys-file "/etc/named.iscdlv.key";

        managed-keys-directory "/var/named/dynamic";

        pid-file "/run/named/named.pid";
        session-keyfile "/run/named/session.key";
};


logging
{
/*      If you want to enable debugging, eg. using the 'rndc trace' command,
 *      named will try to write the 'named.run' file in the $directory (/var/named).
 *      By default, SELinux policy does not allow named to modify the /var/named directory,
 *      so put the default debug log file in data/ :
 */
        /*channel default_debug {
                print-time yes;
                print-category yes;
                print-severity yes;
                file "data/named.run";
                severity dynamic;
        };*/
        channel queries_log {
                file "/var/log/queries" versions 1 size 20m;
                print-time yes;
                print-category yes;
                print-severity yes;
                severity debug 3;
        };

        category queries { queries_log; };
        category client { queries_log;  };
};

zone "." IN {
        type hint;
        file "named.ca";
};

include "/etc/named.rfc1912.zones";
include "/etc/named.root.key";

另请注意，我有一个 quagga riuter 配置为允许通过公共 IP 进行 DNS 解析

/etc/quagga/bgpd.conf

!
! Zebra configuration saved from vty
!   2019/10/11 10:11:45
!
!
router bgp AS
 bgp router-id PublicIP
 network PublicIP/32
 network CoreIP/32
 neighbor DUB1-WGW peer-group
 neighbor DUB1-WGW remote-as AS
 neighbor DUB1-WGW soft-reconfiguration inbound
 neighbor DUB1-WGW route-map XXXXX out
 neighbor CoreBGPIP peer-group DUB1-WGW
 neighbor CoreBGPIP peer-group DUB1-WGW
!
ip prefix-list XXXX seq 5 permit PublicIP/32
ip prefix-list XXXX seq 10 permit PrivateIP/32
!
route-map DNS_TO_GI permit 10
 match ip address prefix-list XXXXX
!
line vty
!

/etc/quagga/zebra.conf

!
! Zebra configuration saved from vty
!   2019/10/11 10:11:45
!
hostname hostname
!
interface ens160
 ipv6 nd suppress-ra
!
interface ens192
 ipv6 nd suppress-ra
!
interface ens192.890
 ipv6 nd suppress-ra
!
interface ens192.892
 ipv6 nd suppress-ra
!
interface XX
 ipv6 nd suppress-ra
!
interface lo
!
ip prefix-list XX seq 5 permit PublicIP3/32
ip prefix-list XX seq 10 permit PrivateIP/32
!
route-map XXXX permit 10
 match ip address prefix-list XXX
!
!
!
line vty
!

# show ip route
Codes: K - kernel route, C - connected, S - static, R - RIP,
       O - OSPF, I - IS-IS, B - BGP, A - Babel,
       > - selected route, * - FIB route

B>* 0.0.0.0/0 [20/10] via neighbor IP, ens192.892, 00:02:18
C>* 127.0.0.0/8 is directly connected, lo
C>* Public IP/32 is directly connected, lo
C>* NeighborSubnet/30 is directly connected, ens192.890
C>* NeighborIP/30 is directly connected, ens192.892
C>* LocalIP/32 is directly connected, lo

我正在使用测试 APN 测试分辨率，虽然当我引入第二个 APN 时我可以将一个 APN 作为 sson 获得分辨率，但我只是在 tcpdump 中遇到以下错误：

11:29:38.065284 IP PublicIP.domain > internal IP.p2pcommunity: 30622 ServFail 0/0/0 (44)
11:29:38.265736 IP PublicIP.domain > internal IP.32209: 12606 ServFail 0/0/0 (37)
11:29:38.266037 IP PublicIP.domain > internal IP.10793: 26678 ServFail 0/0/0 (37)
11:29:38.295727 IP PublicIP.domain > internal IP.ibm_wrless_lan: 23483 ServFail 0/0/0 (33)
11:29:38.296038 IP PublicIP.domain > internal IP.22097: 8347 ServFail 0/0/0 (33)
11:29:38.297532 IP PublicIP.domain > internal IP.31026: 23400 ServFail 0/0/0 (38)
11:29:38.298117 IP PublicIP.domain > internal IP.23707: 26481 ServFail 0/0/0 (38)

并从 /var/log/queries

22-Sep-2020 11:31:07.552 client: debug 3: client InternalIP#61793 (www.facebook.com): error
22-Sep-2020 11:31:07.552 client: debug 3: client InternalIP#61793 (www.facebook.com): send
22-Sep-2020 11:31:07.552 client: debug 3: client InternalIP#61793 (www.facebook.com): sendto
22-Sep-2020 11:31:07.552 client: debug 3: client InternalIP#48008 (2.android.pool.ntp.org): error
22-Sep-2020 11:31:07.552 client: debug 3: client InternalIP#61793 (www.facebook.com): senddone
22-Sep-2020 11:31:07.552 client: debug 3: client InternalIP#61793 (www.facebook.com): next
22-Sep-2020 11:31:07.552 client: debug 3: client InternalIP#61793 (www.facebook.com): endrequest
22-Sep-2020 11:31:07.553 client: debug 3: client InternalIP#48008 (2.android.pool.ntp.org): send
22-Sep-2020 11:31:07.553 client: debug 3: client InternalIP#48008 (2.android.pool.ntp.org): sendto
22-Sep-2020 11:31:07.553 client: debug 3: client InternalIP#48008 (2.android.pool.ntp.org): senddone
22-Sep-2020 11:31:07.553 client: debug 3: client InternalIP#48008 (2.android.pool.ntp.org): next
22-Sep-2020 11:31:07.553 client: debug 3: client InternalIP#48008 (2.android.pool.ntp.org): endrequest

我真的不确定如何解决这个问题，任何指针或建议将不胜感激

dig 命令的输出

dig facebook.com

; <<>> DiG 9.9.4-RedHat-9.9.4-74.el7_6.1 <<>> facebook.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 7204
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4000
;; QUESTION SECTION:
;facebook.com.          IN  A

;; ANSWER SECTION:
facebook.com.       93  IN  A   31.13.86.36

;; Query time: 2 msec
;; SERVER: internal DNS#53(Internal DNS)
;; WHEN: Tue Sep 22 19:38:58 UTC 2020
;; MSG SIZE  rcvd: 57


dig @PublicIP facebook.com

; <<>> DiG 9.9.4-RedHat-9.9.4-74.el7_6.1 <<>> @PublicIP facebook.com
; (1 server found)
;; global options: +cmd
;; connection timed out; no servers could be reached

dig @208.67.222.222 facebook.com

; <<>> DiG 9.9.4-RedHat-9.9.4-74.el7_6.1 <<>> @208.67.222.222 facebook.com
; (1 server found)
;; global options: +cmd
;; connection timed out; no servers could be reached

ip netns exec gi tcpdump -n -f 'port 53' -i any
09:55:35.676645 IP PublicIP.domain > InternalIP.46571: 36451 ServFail 0/0/0 (32)
09:55:35.676939 IP PublicIP.domain > InternalIP.37817: 52592 ServFail 0/0/0 (32)
09:55:35.677865 IP PublicIP.domain > InternalIP41737: 52624 ServFail 0/0/0 (32)
09:55:35.713870 IP PublicIP.34042 > 193.0.14.129.domain: 11264 [1au] A? mtalk.google.com. (45)
09:55:35.713914 IP PublicIP.11218 > 193.0.14.129.domain: 3623 [1au] NS? . (28)
09:55:35.768649 IP 193.0.14.129.domain > PublicIP.11218: 3623*-| 0/0/1 (28)
09:55:35.784456 IP 193.0.14.129.domain > PublicIP.34042: 11264-| 0/0/1 (45)
09:55:36.045130 IP PublicIP.wcbackup > 192.112.36.4.domain: 28368 A? update.googleapis.com. (39)
09:55:36.063323 IP InternalIP.49382 > PublicIP.domain: 57145+ A? accounts.google.com. (37)
09:55:36.064459 IP PublicIP.48169 > 193.0.14.129.domain: 15825 [1au] A? accounts.google.com. (48)
09:55:36.065883 IP APNIP.54312 > PublicIP.domain: 53585+ A? accounts.google.com. (37)
09:55:36.080202 IP 192.112.36.4.domain > PublicIP.wcbackup: 28368- 0/13/14 (499)
09:55:36.120905 IP 193.0.14.129.domain > PublicIP.48169: 15825- 0/15/27 (1182)
09:55:36.170289 IP InternalIP.59759 > PublicIP.domain: 52061+ A? www.google.com. (32)
09:55:36.224316 IP PublicIP.5346 > 192.112.36.4.domain: 40438 A? www.facebook.com. (34)
09:55:36.257993 IP 192.112.36.4.domain > PublicIP.5346: 40438- 0/13/14 (494)
09:55:36.441576 IP PublicIP.domain > InternalIP.65408: 45517 ServFail 0/0/0 (39)
09:55:36.441666 IP PublicIP.domain > InternalIP.60664: 54663 ServFail 0/0/0 (39)
09:55:36.442994 IP PublicIP.domain > InternalIP.48634: 56799 ServFail 0/0/0 (39)
09:55:36.443474 IP PublicIP.domain > InternalIP.36045: 34980 ServFail 0/0/0 (39)

我一直在 centos 7 上构建 DNS 服务器（我继承了这些但需要进行更改以重新利用它们）

当我运行任何 rndc 命令时它不起作用，即当我尝试 rndc reload 时，如果成功，我没有得到确认，甚至没有得到命令失败

另请注意，我使用的是 linux 命名空间，因此典型的 rndc 命令具有前缀“ip entns exec gi”

如果我同时检查 /var/log/messages，与另一台 DNS 服务器相比，我不会得到任何输出。

我在下面发布了必要服务的输出：

/usr/lib/systemd/system/named.service

[Unit]
Description=Berkeley Internet Name Domain (DNS)
Wants=nss-lookup.target
Wants=named-setup-rndc.service
Before=nss-lookup.target
After=network.target
After=named-setup-rndc.service

[Service]
Type=forking
Environment=NAMEDCONF=/etc/named.conf
EnvironmentFile=-/etc/sysconfig/named
Environment=KRB5_KTNAME=/etc/named.keytab
PIDFile=/run/named/named.pid

ExecStartPre=/bin/bash -c 'if [ ! "$DISABLE_ZONE_CHECKING" == "yes" ]; then /usr/sbin/named-checkconf -z "$NAMEDCONF"; else echo "Checking of zone files is disabled"; fi'

ExecStart=/usr/sbin/named -u named -c ${NAMEDCONF} $OPTIONS

ExecReload=/bin/sh -c '/usr/sbin/rndc reload > /dev/null 2>&1 || /bin/kill -HUP $MAINPID'

ExecStop=/bin/sh -c '/usr/sbin/rndc stop > /dev/null 2>&1 || /bin/kill -TERM $MAINPID'

PrivateTmp=true

[Install]
WantedBy=multi-user.target

/usr/lib/systemd/system/named-setup-rndc.service

[Unit]
Description=Generate rndc key for BIND (DNS)

[Service]
Type=oneshot

ExecStart=/usr/libexec/generate-rndc-key.sh

我试图删除 rndc.key 文件并使用命名服务重新创建，但这没有奏效。我真的很难过。有人对此有经验吗

命名后重新启动日志

-- Unit named-setup-rndc.service has begun starting up.
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 systemd[1]: Started Generate rndc key for BIND (DNS).
-- Subject: Unit named-setup-rndc.service has finished start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
-- 
-- Unit named-setup-rndc.service has finished starting up.
-- 
-- The start-up result is done.
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 systemd[1]: Starting Berkeley Internet Name Domain (DNS)...
-- Subject: Unit named.service has begun start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
-- 
-- Unit named.service has begun starting up.
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: starting BIND 9.9.4-RedHat-9.9.4-74.el7_6.1 -u named -c /etc/named.conf -4
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: built with '--build=x86_64-redhat-linux-gnu' '--host=x86_64-redhat-linux-gnu' '-
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: ----------------------------------------------------
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: BIND 9 is maintained by Internet Systems Consortium,
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: Inc. (ISC), a non-profit 501(c)(3) public-benefit
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: corporation.  Support and training for BIND 9 are
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: available at https://www.isc.org/support
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: ----------------------------------------------------
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: adjusted limit on open files from 4096 to 1048576
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: found 8 CPUs, using 8 worker threads
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: using 8 UDP listeners per interface
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: using up to 21000 sockets
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: loading configuration from '/etc/named.conf'
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: reading built-in trusted keys from file '/etc/named.root.key'
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: initializing GeoIP Country (IPv4) (type 1) DB
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: GEO-106FREE 20180327 Build 1 Copyright (c) 2018 MaxMind Inc All Rights Reserved
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: initializing GeoIP Country (IPv6) (type 12) DB
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: GEO-106FREE 20180605 Build 1 Copyright (c) 2018 MaxMind Inc All Rights Reserved
Sep 24 14:01:40 DUB1-00-VM-GI_DNS1 named[12637]: GeoIP City (IPv4) (type 2) DB not available

我也相信调试日志被设置为写入 /var/named/data/named.run 下面发布的输出

tail -f /var/named/data/named.run
all zones loaded
running
shutting down
stopping command channel on 127.0.0.1#953
stopping command channel on ::1#953
no longer listening on 127.0.0.1#53
exiting
managed-keys-zone: loaded serial 1021
all zones loaded
running