一個MSS參數引發的“血案”

  最近在玩一些淘汰下來的FW，在馬云家淘了一些二手的玩玩，在家搭建了一臺zabbix監控，配置了onealert的免費通知插件（支持微信、QQ、郵件、短信、電話等），用來監控我家小PP看動畫片時長，時間過長就要遠程斷網或shutdown交換機接口，因為當著面關他電視后果很嚴重，斷他網他會知道是“壞了”，沒那么鬧騰。

  回到正題，以前一直用無線路由器做NAT轉發，發現即使是cisco 6900和網件R 7000等千元路由器級別都會用到死機。后來幫別人做項目發現juniper ssg和SRX這種企業級的FW在某寶只要幾百元，果斷出手搞了一些不同型號來測試。

本文的主角：JUNIPER SRX 210H正式登場

當我用210配置完PPPOE后，部分網站可以打開，部分網站打不開，并且在JUNIPER SSG5上面沒有這個問題，所以斷定問題在210上。排錯思路如下：

一、檢查PPPOE鏈路狀態

看起來正常

admin@YY-SRX100H#run show interfaces pp0

Physical interface: pp0, Enabled, Physical link is Up

  Interface index: 128, SNMP ifIndex: 501

  Type: PPPoE, Link-level type: PPPoE, MTU: 1532

  Device flags   : Present Running

  Interface flags: Point-To-Point SNMP-Traps

  Link type      : Full-Duplex

  Link flags     : None

  Input rate     : 232 bps (0 pps)

  Output rate    : 0 bps (0 pps)

  Logical interface pp0.0 (Index 79) (SNMP ifIndex 563)

    Flags: Point-To-Point SNMP-Traps 0x0 Encapsulation: PPPoE

    PPPoE:

      State: SessionUp, Session ID: 34772,

      Session AC name: SZ-BJ-BAS-5.MAN.NE40E, Remote MAC address: da:86:8e:6c:00:19,

      Configured AC name: None, Service name: None,

      Auto-reconnect timeout: 10 seconds, Idle timeout: Never,

      Underlying interface: fe-0/0/1.0 (Index 78)

    Input packets : 24 

    Output packets: 16

  Keepalive settings: Interval 10 seconds, Up-count 1, Down-count 3

  Keepalive: Input: 3 (00:00:08 ago), Output: 7 (00:00:01 ago)

  LCP state: Opened

  NCP state: inet: Opened, inet6: Not-configured, iso: Not-configured, mpls: Not-configured

  CHAP state: Closed

  PAP state: Success

    Security: Zone: Null

    Protocol inet, MTU: 1492

      Flags: Sendbcast-pkt-to-re, User-MTU, Negotiate-Address

      Addresses, Flags: Kernel Is-Preferred Is-Primary

        Destination: 183.12.26.1, Local: 183.12.26.79

二、檢查區域和策略

也都正常，策略全放開

三、根據網上的建議調整MTU為1400

然并卵，問題依舊

set interfaces pp0 unit 0 family inet mtu 1400

四、根據度娘搜遍了大量相關的蛛絲馬跡，發現一個很少有人問津的tcp-mss參數調整

憑借我多年運維的經驗直覺告訴我，真相很快就要浮出水面了。

The maximum segment size (MSS) is a parameter of the options field of the TCP header that specifies the largest amount of data, specified in bytes, that a computer or communications device can receive in a single TCP segment. It does not count the TCP header or the IP header.^[1] The IP datagram containing a TCP segment may be self-contained within a single packet, or it may be reconstructed from several fragmented pieces; either way, the MSS limit applies to the total amount of data contained in the final, reconstructed TCP segment.

To avoid fragmentation in the IP layer, a host must specify the maximum segment size as equal to the largest IP datagram that the host can handle minus the IP header size and TCP header sizes.^[2] Therefore, IPv4 hosts are required to be able to handle an MSS of 536 octets (= 576^[3] - 20 - 20) and IPv6 hosts are required to be able to handle an MSS of 1220 octets (= 1280^[4] - 40 - 20).

Small MSS values will reduce or eliminate IP fragmentation, but will result in higher overhead.^[5]

Each direction of data flow can use a different MSS.

For most computer users, the MSS option is established by the operating system.

上面一段話其實簡要概之就是，它和TCP有關。。。也別太較真了

于是乎就抱著試一試的態度，結果之前打不開的網頁都能打開了

set security flow tcp-mss all-tcp mss 1350

五、pppoe全部配置參考本人以下博文

http://yangye.blog.51cto.com/922715/1874180