Debian 8.6上的Ghost NTP服务器

因此，我和大学的IT安全团队一直在不停地走来走去...任何人对此都有任何想法：

我最近为我的实验室设置了一台小型文件服务器，在专用计算机上运行Debian 8.6（英特尔Avoton C2550处理器，很乐意在需要时提供更多硬件信息，但我认为没有必要）。Debian的安装没有任何问题，当时我还安装了Samba，NTP，ZFS和python。看起来一切正常，所以我让它在实验室的角落坐了几个星期。

大约两周前，我收到了IT团队的电子邮件，其中说我的服务器已“受损”，并且易受NTP放大/ DDoS攻击（使用CVE-2013-5211的NTP放大攻击，如https中所述）使用： //www.us-cert.gov/ncas/alerts/TA14-013A）。他们指向的标志是端口123上有很多NTPv2流量。奇怪的是，他们标识为（*.*.*.233）的IP地址与为服务器配置并通过ifconfig（*.*.*.77）报告的IP地址不同。但是，一些基本的故障排除表明我的计算机确实在端口123上生成了此流量（如tcpdump所示）。

这就是怪异开始的地方。我首先浏览了为CVE-2013-5211推荐的“修复程序”（同时更新了版本4.2.7之后的NTP并禁用了monlist功能）。两者均未阻止交通流量。然后，我尝试通过IP表阻止UDP 123端口：

$ /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

但这对流量也没有影响。我最终尝试从系统中清除NTP，但这对流量也没有影响。截至今天下午，nmap仍在报告：

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

由于NTP已从系统中清除了几周，所以这一切都很奇怪。

在这条道路上走到尽头之后，我开始考虑整个IP地址不匹配的问题。我的计算机似乎同时位于* .233和* .77 IP上（已通过在连接有以太网电缆的情况下成功ping通，并且在拔出电缆后都无法ping通），但是* .233从未出现在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中没有引用* .233，因此我看不到此IP分配来自何处。

因此，我有两个可能相关的问题，希望有人能为我提供帮助：1）如何消除从服务器喷涌而来的NTP流量，以使IT摆脱困境？2）我的服务器正在使用的第二个IP地址是怎么回事，如何删除它？

谢谢，伙计们:)

更新：根据要求： $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：我没有提到，除了IP地址不匹配外，MAC ID也不匹配。这确实使我对流量是否确实来自我的计算机进行了三思。但是：（1）从网络上拔出服务器后，流量消失了；（2）移至其他网络端口，流量随之而来；（3）tcpdump port 123显示流量异常：

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

（即无）

$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：这些数据包是几天前的典型数据。今天（但是，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不匹配我的硬件MAC地址是：d0：50：99：51：78：7a流量与MAC相关联：bc：5f：f4：fe：a1：00

更新5：根据要求，对* .233进行端口扫描：

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)

并在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)

— 蒂姆·奥奇（Tim Otchy）
source

听起来您的机器已经受到威胁，并试图隐藏这一事实。iptables -L -v -n和的完整输出是什么ip addr ls。

— Mark Wagner

@TimOtchy需要说明的是，当断开与服务器的以太网电缆连接时，是在服务器背面还是在交换机上进行操作？您是否有备用交换机，并且可以将以太网电缆从服务器连接到交换机，但没有其他（电源除外）插入交换机。想法是在服务器上启用链接，但以其他方式隔离该链接，然后查看* 233是否可访问。

— icarus

在给定（server-> switch no_connection LAN）的情况下，您可以从服务器ping通和（server no_connection LAN），并且只能从服务器ping * .77，我认为我们可以得出结论，服务器也正在服务* .233。下一个问题，“服务器”是“大”框吗？我认为它可能具有单独的机箱管理CPU，或者它是x86设备并具有内置的监视设备。我倾向于对* .233运行完整的端口扫描，看看打开了哪些端口。

— icarus

请参阅en.wikipedia.org/wiki/Intelligent_Platform_Management_Interface，这是一个独立的CPU，与主CPU，BIOS或主OS无关。

— icarus

看起来该修复程序应该针对正在运行ssh（以及nmap猜测的Linux）的BMC（IPMI）处理器。Asrock网站上已安装了自2016年3月23日起的软件，但我想那无济于事。我的想法是查看是否可以使用* 233地址。我猜想您可能需要在BMC选项下的BIOS设置中设置用户名和密码？？？

— icarus

这是带有IPMI的服务器类计算机。导致问题的“ ghost” NTP服务器在系统的BMC处理器而非主CPU上运行。

— 伊卡洛斯
source

刚通过24小时，就没有来自此mac / IP的任何新NTP流量。绝对是BMC处理器运行的固件版本非常旧，并且使用NTP程序包仍然容易受到放大漏洞的攻击。感谢您为解决此问题提供的所有帮助！

— 蒂姆·奥奇

如前所述，您的IPMI BMC可能是问题所在。如果无法访问IPMI界面的Web界面或ssh界面，则可以尝试在Debian安装中使用IPMI工具获得访问权限：

apt install ipmitool

安装完成后，您可以像这样将命令传递给BMC（如果它在端口1上）：

ipmitool lan set 1 ipaddr 192.168.1.211

这是使用IPMITool进行IPMI网络配置的资源。 man ipmitool如果被卡住，总是一本好书...

如果需要，此页面包含有关重置BMC的信息。

祝好运！

— 问号
source