1

I am in the progress of doing ISSU upgrades on some nexus 7710 chassis' with dual supervisor modules. I set a ping test that to traverse across the switch shortly before the upgrade and experienced about 15-20 seconds of lost pings. Is this test scenario a good one and does this amount of downtime seem acceptable? Is cisco's marketing on this feature just bs? I see the same behavior for a supervisor switchover as well.

2

The best tip I can give for anything involving ISSU on 7K's is to read- and re-read- the release notes. There are very specific versions that are supported as origin/destination and there are also some features that are going to work better with ISSU than others. The other caveat to consider is that in some instances there may be requirements for firmware or EPLD upgrades that could create a brief outage (again - usually mentioned in release notes).

In general I can tell you that the ISSU process can be truly hitless but that you need to stick within the limits of what QA has specifically validated.

| improve this answer | |
  • I was able to check the release notes and the path we are using is supported. I'm not certain this is an ISSU issue but rather the failing over of the HA pair of supervisors. – stets Feb 15 '17 at 19:59
  • What version to what version? All F3 LC's or do you have some F2e or M3's in the mix? Also - which features are in use? It may be that there are one or two specific things causing the issue. There tend to be improvements to address specific features in ISSU - so a recently implemented set like VXLAN might cause an issue where older features like VPC or ASM multicast might not. – rnxrx Feb 16 '17 at 3:33
  • 7.2.1.D1.1 to 7.3.1.D1.1. Just the dual sup and 2 F3 cards. Features are -feature privilege feature scp-server feature sftp-server feature tacacs+ feature bgp feature eigrp feature udld feature interface-vlan feature netflow feature hsrp feature lacp feature vpc feature lldp feature bfd feature sla sender feature sla responder – stets Feb 16 '17 at 14:59
1

I found the issue. The interfaces of the 7K we were upgrading didn't have the 'no ip redirects' command on. This caused a drop in BFD between the neighboring router we were testing from and also dropped our EIGRP neighborship. This list has similar information: https://lists.gt.net/cisco/nsp/141379

| improve this answer | |
  • You should accept your answer so that the question doesn't keep popping up forever, looking for an answer. – Ron Maupin Feb 17 '17 at 14:44
  • Will do, stack exchange has set a limit that I can't accept it for another 3 hours. – stets Feb 17 '17 at 15:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.