the topology refer to attachment.S9300 run RRPP as free loop and fast convergence protocol as well as CX600 run VPLS point to point to close the ring. services vlans as RRPP protect vlans connect to CX600 binding to VSI,implement the VPLS over TE terminatead the vlans in MSC central site's CE.
phenomena description :
TE configure as Hot-standby mode,the primary and backup paths use explicit path, the information as below:
path1. primary path from 203 to MSC_Minsk_1.
path2. backup path from 203 to MSC_Minsk_1.
path3 best-effort path from MSC_Minsk_1(auto caculate by rsvp).
when the link from 203 to 261 is down,all sevices connect to 203 interrupted.
1.when the service interrupation,203 still can be managed,mean IGP is normal.
2.login some available s93s and try to ping the other S93s on the same ring,and found can reach these S93s,so confirmed ARP and MAC refresh work.
3.start to shutdown the port in CX600 which connect to S93 rings to force the RRPP mater node change the direction and found the service recover. so can confirm is not the rrpp protocol.
4.check the VPLS down and TE down as well.
Due to the link from 226 to MSC-1 had been down several days ago.that mean the Backup path is down before the primary path.when the link from 203 to 261 is down,the primary path wend down as well.so the Tunnel will try to generate the best-effort path to transfer the traffic.but the service unfortunately,the Tunnel from MSC-1 to 203 did not be configured best-offort function correctly.
1.IGP routing table.
2. MAC address did not refresh on time.
3.RRPP in S93 work not correctly.
4.VPLS or TE down.