RE and swtiching control board redundancy

1.Understanding Routing Engine Redundancy on Juniper Networks Routers 1.1.RE redundancy overview

GRE未配置-备用RE重启交换平面，PFE被初始化，流量中断。

GRE配置-interface and kernel info is preserved。重启RPD。所有硬件和端口热重启

GRE+NSR配置-流量不中断，interface kernel and routing protocol info preserve。特定协议有问题

GRE+graceful restart-流量不中断，interface kernel info preserve。从邻居路由器收集和还原路由信息

1.2.conditions that trigger RE failover

1.hardware failure,同时可以配置hard disk错误进行RE切换，需配置failover on-disk-failure.

2.software failure.such as kernel crash or a cpu lock.cofig failover on-loss-of-keepalives to enable backup re to take mastership when it dectect a loss of keepalive signal.

3.specific software prcess fails.cofig failover other-routing-engine to make backup re take mastership when on or more specific prcesses fail at least 4 times within 30s.

1.3.default re redundancy behavior

by default,re0 is master

1.4.re redundancy on a tx matrix router

you must run the same junos release on all re.or will lead re switch,cant't switch

1.5.situations that requiure you to halt re

1.if you need remove the master re,you must first halt the bakcup re,you can use request system halt both-routing-eng

2.to shut down only the backup re,issue the request system halt on backup re.

3.if you halt the master re and do not power it off or remove it,the bakcup re remain inactive unless you have configured failover on-loss-keepalive

2.switching control board redundancy 2.1.redundancy CFEB on M10i

the cfeb perform the following functions:

1.route lkup,perform route lkup using the forwarding table stored in ×××AM

2.management of shared memory.

3.transfer of outgoing data packets.pass packet to the destionation fic or pic

4.transfer of exception and control packets.pass to CFEB,remainder are sent to RE for further processing.errors originating in PFE are sent to the RE.

M10i has 2 CFEB

2.2.redundancy FEBs on M120

1.M120 supports up to 6 FEB

2.FPCs are separate from the FEB,which handle packet forwarding

3.use fcp-feb-connectivity to map a FPC to FEB(this FEB can't config as a backup).by default,FPC assigned automatically to the FEB with the same slot NO.

4.you can config a FEB as backup for one or more FEB by config a FEB redundancy group,but a FEB can belong to only one group.

5.you can disable automatic failover for any redundancy group by including the no-auto-failover

6.issue the request chassis redundancy feb slot X switch-to-backup to manual switchover.and xxxx revert-from-bakcup to active previously active FEB.

7.modify the configuration for an existing group so that a FEB connect to a different FPC,the FEB will reboot.

2.3.redundancy SSBs on the M20

M20 hold 2 SSB,

function:

1.shared memory management on the FPC

2.pass data cell to FPC

3.forwarding table route lookup.

4.system component monitor,like temperature.

5.transfer exception and control packet.SSB processes almost all of them,remain sent to RE.

6.FPC reset control.if it dectects errors in an FPC,the SSB reset the FPC,after 3 unsuccessful resets,the SSB take the FPC offline

2.4.redundancy SFMs on M40e and M160

SFM(switching and forwarding module) provide route lookup,filtering and switching

M40e hold up to 2 SFM

if the active SFM fail or removed from the chassis,forwarding halt until sthandby SFM boots and active,it takes approximately 1m,sys route config can take additional time

M160 hold up to 4 SFMs

all SFMs active at the same time