In order to increase the capacity of Private Cloud infrastructure we plan to replace 2 new cards on each Nexus 7000.
In this task, we will need to move the links between 2 Switches and also to the routers.
The task is scheduled on the night of Wednesday, October the 3th to Thursday, October the 4th. (00:00 CEST)
Update(s):
Date: 2012-10-03 23:49:53 UTC All links are switched of the cards M1 to M2 and the vPC is re-established through new cards. Both switches are performing well. We are removing the cards M1 of the chassis.
Date: 2012-10-03 23:05:42 UTC uplinks are switched on pcc-1b
pcc-1a in progress
Date: 2012-10-03 23:05:21 UTC The vPC is re-established. We have noticed an interruption of traffic
Date: 2012-10-03 23:03:44 UTC We are reactivating the peer-link on the new cards M2.
Date: 2012-10-03 22:32:47 UTC The link between Nexus is down. We are switching on the new cards.
Date: 2012-10-03 22:32:03 UTC pcc-1b-n7# 2012 Oct 4 00:25:49 pcc-1b-n7 %$ VDC-1 %$ %VPC-2-VPC_SUSP_ALL_VPC: Peer-link going down,
Date: 2012-10-03 22:27:16 UTC We are starting migration.
We are cutting off the \"peer link\" between both Nexus 7000.
Date: 2012-10-03 22:26:07 UTC Cards are running on pcc-1b-n7.
Date: 2012-10-03 22:25:14 UTC Both M2 cards are online.
2012 Oct 3 23:20:15 pcc-1a-n7 %SYSMGR-SLOT7-5-MODULE_ONLINE: System Manager has received notification of local module becoming online.
2012 Oct 3 23:20:16 pcc-1a-n7 %PLATFORM-5-MOD_STATUS: Module 7 current-status is MOD_STATUS_ONLINE/OK
2012 Oct 3 23:20:15 pcc-1a-n7 %SYSMGR-SLOT8-5-MODULE_ONLINE: System Manager has received notification of local module becoming online.
2012 Oct 3 23:20:16 pcc-1a-n7 %PLATFORM-5-MOD_STATUS: Module 8 current-status is MOD_STATUS_ONLINE/OK
We ar inserting cards in pcc-1b-n7.
Date: 2012-10-03 22:24:50 UTC We are inserting M2 cards in pcc-1a-n7.
Date: 2012-10-03 14:16:44 UTC Details in relation with tonight work.
The purpose of the intervention is to migrate \"M1\" cards on which uplinks are connected and \"peer-links\" between the two switches to the next \"M2\" generation. This is in order to continue to evolve the platform.
We worked closely with the inges of Cisco to develop the exact procedure of migration. This procedure has been validated by Cisco in lab using traffic generators. During these lab tests, the \"traffic loss\" was 766ms during the migration phase of peer-links and 1.3 s during those of the migration of uplinks to routers (1.3s for each of the two routers then). The duration of the \"traffic loss\" being the sum of cuts of traffic in the uplink direction (from the pcc infra to routers) and descending (the reverse). Our goal will be to stick closer to times measured in lab.