Debian 8.6 上的 Ghost NTP 伺服器

因此，我和大學 IT 安全團隊一直在不停地討論這個問題……任何人對此都有任何想法：

我最近為我的實驗室設置了一個小型文件伺服器，在專用電腦上執行 Debian 8.6（Intel Avoton C2550 處理器——如果需要，很樂意提供更多硬體資訊，但我認為沒有必要）。Debian安裝沒有任何問題，當時我還安裝了Samba、NTP、ZFS和python。一切似乎都很好，所以我讓它在實驗室的角落裡執行了幾個星期。

大約兩週前，我收到了一封來自 IT 團隊的電子郵件，說我的伺服器已經“受到威脅”並且很容易被用於 NTP 放大/DDoS 攻擊（使用 CVE-2013-5211 的 NTP 放大攻擊，如https 中所述： //www.us-cert.gov/ncas/alerts/TA14-013A）。他們指出的標誌是埠 123 上有大量 NTPv2 流量。奇怪的是，他們辨識出的來自 ( *.*.*.233) 的 IP 地址與我的伺服器配置並通過 ifconfig ( *.*.*.77) 報告的 IP 地址不同。儘管如此，一些基本的故障排除顯示我的電腦確實在埠 123 上產生了這種流量（正如 tcpdump 所揭示的那樣）。

這就是奇怪的開始。我首先瀏覽了針對 CVE-2013-5211 推薦的“修復程序”（更新 NTP 超過 4.2.7 版以及禁用 monlist 功能）。兩者都沒有阻止交通流量。然後我嘗試通過 IP 表阻止 UDP 123 埠：

$ /sbin/iptables -A INPUT -o eth0 -p udp --destination-port 123 -j DROP
$ /sbin/iptables -A OUTPUT -o eth0 -p udp --destination-port 123 -j DROP

但這也對交通沒有影響。我終於嘗試從系統中清除 NTP，但這對流量也沒有影響。截至今天下午，nmap 仍在報告：

Starting Nmap 5.51 ( http://nmap.org ) at 2016-12-19 16:15 EST
Nmap scan report for *.233
Host is up (0.0010s latency).
PORT    STATE SERVICE
123/udp open  ntp
| ntp-monlist:
|   Public Servers (2)
|       50.116.52.97    132.163.4.101
|   Public Clients (39)
|       54.90.159.15    185.35.62.119   185.35.62.233   185.35.63.86
|       54.197.89.98    185.35.62.142   185.35.62.250   185.35.63.108
|       128.197.24.176  185.35.62.144   185.35.62.251   185.35.63.128
|       180.97.106.37   185.35.62.152   185.35.63.15    185.35.63.145
|       185.35.62.27    185.35.62.159   185.35.63.27    185.35.63.146
|       185.35.62.52    185.35.62.176   185.35.63.30    185.35.63.167
|       185.35.62.65    185.35.62.186   185.35.63.34    185.35.63.180
|       185.35.62.97    185.35.62.194   185.35.63.38    185.35.63.183
|       185.35.62.106   185.35.62.209   185.35.63.39    185.35.63.185
|_      185.35.62.117   185.35.62.212   185.35.63.43

這很奇怪，因為 NTP 已經從系統中清除了數週了。

在這條路走上死胡同後，我開始思考整個 IP 地址不匹配的問題。我的電腦似乎同時使用了 *.233 和 *.77 IP（通過連接乙太網電纜成功 ping 並且拔下電纜均不可用），但 *.233 從未出現在 ifconfig 中：

Link encap:Ethernet  HWaddr d0:XX:XX:51:78:XX  
inet addr:*.77  Bcast:*.255  Mask:255.255.255.0
inet6 addr: X::X:X:X:787a/64 Scope:Link
UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
RX packets:23023571 errors:0 dropped:1362 overruns:0 frame:0
TX packets:364849 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000 
RX bytes:7441732389 (6.9 GiB)  TX bytes:44699444 (42.6 MiB)
Memory:df300000-df37ffff 

/etc/network/interfaces 中沒有對 *.233 的引用，所以我看不到這個 IP 分配來自哪裡。

所以，我有兩個可能相關的問題，希望有人能幫助我：1）我怎樣才能消除從我的伺服器噴出的這種 NTP 流量以讓 IT 擺脫我的支持？2) 我的伺服器所在的第二個 IP 地址是怎麼回事，我該如何刪除它？

謝謝，伙計們:)

更新：根據要求： $iptables -L -v -n

Chain INPUT (policy ACCEPT 57 packets, 6540 bytes)
pkts bytes target     prot opt in     out     source               destination         

Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
pkts bytes target     prot opt in     out     source               destination         

Chain OUTPUT (policy ACCEPT 27 packets, 2076 bytes)
pkts bytes target     prot opt in     out     source               destination

和$ip addr ls

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default 
   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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
   link/ether d0:50:99:51:78:7a brd ff:ff:ff:ff:ff:ff
   inet *.77/24 brd *.255 scope global eth0
      valid_lft forever preferred_lft forever
   inet *.167/24 brd *.255 scope global secondary dynamic eth0
      valid_lft 24612sec preferred_lft 24612sec
   inet6 X::X:X:X:787a/64 scope link 
      valid_lft forever preferred_lft forever
3: eth1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000
   link/ether d0:50:99:51:78:7b brd ff:ff:ff:ff:ff:ff

更新2：我沒有提到除了IP地址不匹配外，MAC ID也不匹配。這真的讓我三思而後行，流量是否確實來自我的機器。但是：（1）從網路上拔下我的伺服器使流量消失了；(2) 移動到不同的網路埠，流量隨之而來；(3)tcpdump port 123顯示異常流量：

13:24:33.329514 IP cumm024-0701-dhcp-233.bu.edu.ntp > 183.61.254.77.44300: NTPv2, Reserved, length 440
13:24:33.329666 IP cumm024-0701-dhcp-233.bu.edu.ntp > 183.61.254.77.44300: NTPv2, Reserved, length 440
13:24:33.329777 IP cumm024-0701-dhcp-233.bu.edu.ntp > 183.61.254.77.44300: NTPv2, Reserved, length 296

更新 3： $ss -uapn 'sport = :123'

State      Recv-Q Send-Q                  Local Address:Port                    Peer Address:Port 

（即，什麼都沒有）

$sudo cat /proc/net/dev

Inter-|   Receive                                                |  Transmit
face |bytes    packets errs drop fifo frame compressed multicast|bytes    packets errs drop fifo colls carrier compressed
   lo:  327357    5455    0    0    0     0          0         0   327357    5455    0    0    0     0       0          0
 eth1:       0       0    0    0    0     0          0         0        0       0    0    0    0     0       0          0
 eth0: 13642399917 36270491    0 6522    0     0          0   2721337 45098276  368537    0    0    0     0       0          0

更新 4：這些數據包是幾天前的典型數據。今天（但是是的，仍然很高）：

20:19:37.011762 IP cumm024-0701-dhcp-233.bu.edu.ntp > 103.56.63.147.26656: NTPv2, Reserved, length 152
20:19:37.011900 IP cumm024-0701-dhcp-233.bu.edu.ntp > 202.83.122.78.58066: NTPv2, Reserved, length 152
20:19:37.012036 IP cumm024-0701-dhcp-233.bu.edu.ntp > 103.56.63.147.17665: NTPv2, Reserved, length 152
20:19:37.014539 IP cumm024-0701-dhcp-233.bu.edu.ntp > 202.83.122.78.27945: NTPv2, Reserved, length 152
20:19:37.015482 IP cumm024-0701-dhcp-233.bu.edu.ntp > 202.83.122.78.42426: NTPv2, Reserved, length 152
20:19:37.015644 IP cumm024-0701-dhcp-233.bu.edu.ntp > 103.56.63.147.16086: NTPv2, Reserved, length 152

$ sudo ss -uapn '( sport = :42426 or dport = :42426 )'
State       Recv-Q Send-Q                                                        Local Address:Port                                                          Peer Address:Port 

是的，我可以 ping *.233 IP：

$ping 128.197.112.233
PING 128.197.112.233 (128.197.112.233) 56(84) bytes of data.
64 bytes from 128.197.112.233: icmp_seq=1 ttl=64 time=0.278 ms
64 bytes from 128.197.112.233: icmp_seq=2 ttl=64 time=0.282 ms
64 bytes from 128.197.112.233: icmp_seq=3 ttl=64 time=0.320 ms

不，MAC 不匹配 我的硬體 MAC 地址是：d0:50:99:51:78:7a 流量與 MAC 相關聯：bc:5f:f4:fe:a1:00

更新 5：根據要求，對 *.233 進行埠掃描：

Starting Nmap 6.00 ( http://nmap.org ) at 2016-12-20 20:38 EET
NSE: Loaded 17 scripts for scanning.
Initiating SYN Stealth Scan at 20:38
Scanning cumm024-0701-dhcp-233.bu.edu (128.197.112.233) [1024 ports]
Discovered open port 22/tcp on 128.197.112.233
Completed SYN Stealth Scan at 20:38, 9.79s elapsed (1024 total ports)
Initiating Service scan at 20:38
Scanning 1 service on cumm024-0701-dhcp-233.bu.edu (128.197.112.233)
Completed Service scan at 20:38, 0.37s elapsed (1 service on 1 host)
Initiating OS detection (try #1) against cumm024-0701-dhcp-233.bu.edu (128.197.112.233)
Initiating Traceroute at 20:38
Completed Traceroute at 20:38, 0.10s elapsed
NSE: Script scanning 128.197.112.233.

[+] Nmap scan report for cumm024-0701-dhcp-233.bu.edu (128.197.112.233)
Host is up (0.083s latency).
Not shown: 1013 filtered ports

PORT    STATE  SERVICE VERSION
21/tcp  closed ftp
22/tcp  open   ssh     OpenSSH 5.5p1 Debian 6+squeeze1 (protocol 2.0)
23/tcp  closed telnet
25/tcp  closed smtp
43/tcp  closed whois
80/tcp  closed http
105/tcp closed unknown
113/tcp closed ident
210/tcp closed z39.50
443/tcp closed https
554/tcp closed rtsp

Device type: general purpose
Running: Linux 2.6.X
OS CPE: cpe:/o:linux:kernel:2.6
OS details: DD-WRT v24-sp2 (Linux 2.6.19)
Uptime guess: 45.708 days (since Sat Nov  5 03:39:36 2016)
Network Distance: 9 hops
TCP Sequence Prediction: Difficulty=204 (Good luck!)
IP ID Sequence Generation: All zeros
Service Info: OS: Linux; CPE: cpe:/o:linux:kernel


TRACEROUTE (using port 25/tcp)
HOP RTT      ADDRESS
1   0.95 ms  router1-lon.linode.com (212.111.33.229)
2   0.70 ms  109.74.207.0
3   1.09 ms  be4464.ccr21.lon01.atlas.cogentco.com (204.68.252.85)
4   1.00 ms  be2871.ccr42.lon13.atlas.cogentco.com (154.54.58.185)
5   63.45 ms be2983.ccr22.bos01.atlas.cogentco.com (154.54.1.178)
6   63.60 ms TrusteesOfBostonUniversity.demarc.cogentco.com (38.112.23.118)
7   63.55 ms comm595-core-res01-gi2-3-cumm111-bdr-gw01-gi1-2.bu.edu (128.197.254.125)
8   63.61 ms cumm024-dist-aca01-gi5-2-comm595-core-aca01-gi2-2.bu.edu (128.197.254.206)
9   90.28 ms cumm024-0701-dhcp-233.bu.edu (128.197.112.233)

OS and Service detection performed. Please report any incorrect results at http://nmap.org/submit/ .

Nmap done: 1 IP address (1 host up) scanned in 20.73 seconds
          Raw packets sent: 557 (25.462KB) | Rcvd: 97 (8.560KB)

在 UDP 上：

Starting Nmap 6.00 ( http://nmap.org ) at 2016-12-20 20:44 EET
NSE: Loaded 17 scripts for scanning.
Initiating Ping Scan at 20:44
Scanning 128.197.112.233 [4 ports]
Completed Ping Scan at 20:44, 1.10s elapsed (1 total hosts)
Initiating UDP Scan at 20:44
Scanning cumm024-0701-dhcp-233.bu.edu (128.197.112.233) [1024 ports]
Completed UDP Scan at 20:44, 6.31s elapsed (1024 total ports)
Initiating Service scan at 20:44
Scanning 1024 services on cumm024-0701-dhcp-233.bu.edu (128.197.112.233)
Service scan Timing: About 0.39% done
Service scan Timing: About 3.12% done; ETC: 22:12 (1:25:46 remaining)
Service scan Timing: About 6.05% done; ETC: 21:53 (1:04:39 remaining)
Service scan Timing: About 8.98% done; ETC: 21:46 (0:56:03 remaining)
Discovered open port 123/udp on 128.197.112.233
Discovered open|filtered port 123/udp on cumm024-0701-dhcp-233.bu.edu (128.197.112.233) is actually open
Completed Service scan at 21:31, 2833.50s elapsed (1024 services on 1 host)
Initiating OS detection (try #1) against cumm024-0701-dhcp-233.bu.edu (128.197.112.233)
Retrying OS detection (try #2) against cumm024-0701-dhcp-233.bu.edu (128.197.112.233)
NSE: Script scanning 128.197.112.233.
Initiating NSE at 21:31
Completed NSE at 21:31, 10.02s elapsed

[+] Nmap scan report for cumm024-0701-dhcp-233.bu.edu (128.197.112.233)
Host is up (0.089s latency).
Not shown: 1023 open|filtered ports

PORT    STATE SERVICE VERSION
123/udp open  ntp?

1 service unrecognized despite returning data. If you know the service/version, please submit the following fingerprint at http://www.insecure.org/cgi-bin/servicefp-submit.cgi :
SF-Port123-UDP:V=6.00%I=7%D=12/20%Time=58597D5C%P=x86_64-unknown-linux-gnu
SF:%r(NTPRequest,30,"\xe4\x02\x04\xee\0\0\x8a\xff\0:t\xd9\x84\xa3\x04e\xdb
SF:\xcaeEX\xdbC'\xc5O#Kq\xb1R\xf3\xdc\x03\xfb\xb8\+>U\xab\xdc\x03\xfb\xb8\
SF:+T\xd1\xe9")%r(Citrix,C,"\xde\xc0\x010\x02\0\xa8\xe3\0\0\0\0");

Too many fingerprints match this host to give specific OS details
Network Distance: 9 hops

OS and Service detection performed. Please report any incorrect results at http://nmap.org/submit/ .

Nmap done: 1 IP address (1 host up) scanned in 2863.89 seconds
          Raw packets sent: 175 (6.720KB) | Rcvd: 50 (10.088KB)

這是帶有 IPMI 的伺服器類機器。導致問題的“幽靈”NTP 伺服器在系統的 BMC 處理器上執行，而不是在主 CPU 上執行。

正如已經說過的，您的 IPMI BMC（可能）是問題所在。如果您無法訪問 IPMI 界面的 Web 界面或 ssh 界面，您可以嘗試在您的 Debian 安裝上使用 IPMI Tool 獲得訪問權限：

apt install ipmitool

安裝後，您可以像這樣向 BMC 傳遞命令（如果它在埠 1 上）：

ipmitool lan set 1 ipaddr 192.168.1.211

這是使用 IPMITool 進行 IPMI 網路配置的資源。 man ipmitool如果您遇到困難，這總是一本好書……

如果需要，您可以重置 BMC 。

引用自：https://unix.stackexchange.com/questions/331521