BGP issue in Extreme X770 switch

  • 0
  • 1
  • Question
  • Updated 2 years ago
  • Answered
What do the CLI mean? "configure bgp restart both" and "configure bgp as-display-format asplain"?

There is a switch related issue in live network where we see fail-over from primary to secondary switch and back is not as working expected. Both switches are in ISC MLAG cluster and run BGP peering with other adjacent nodes. The fail-over works when first switch reboots but when the second switch comes back up after reboot, all the iBGP sessions reset/re-establish.
Photo of Sumanta Ghosh

Sumanta Ghosh

  • 458 Points 250 badge 2x thumb

Posted 2 years ago

  • 0
  • 1
Photo of Paul Russo

Paul Russo, Alum

  • 9,694 Points 5k badge 2x thumb
Hello Sumanta

Are you using OSPF IGP for your iBGP?

Thanks
P
Photo of Henrique

Henrique, Employee

  • 10,342 Points 10k badge 2x thumb
Hi Sumanta, regarding the commands:

configure bgp restart both is used for BGP Graceful Restart (both means planned and unplanned). "Act as a graceful BGP restart router for both planned and unplanned restarts."

configure bgp as-display-format asplain "Configures the AS number format displayed in show commands"

  asdot           Display AS numbers in dot notation e.g. 1.10

  asplain         Display AS numbers as plain numbers e.g. 65546

Photo of Paul Russo

Paul Russo, Alum

  • 9,694 Points 5k badge 2x thumb
Hello Sumanta.

Henrigue pointed out what I was fishing for.  If you are using OSPF and the router goes down the failover will be fast however when a OSPF router comes back online it needs to join its adjacency with its neighbors will go down then back up to redo the link.  We do support Graceful OSPF which takes care of that.  Essentially it tells the neighbor router that I am going down and when I come back don't redo the connection.  Simplified :)

That may be what you are seeing the drop in the iBGP may be an issue if you are not using graceful restart on the OSPF protocol.

Thanks
P
Photo of Sumanta Ghosh

Sumanta Ghosh

  • 458 Points 250 badge 2x thumb
Hi Paul

Thanks for your reply. I have OSPF and BGP running, but for different purpose. The issue happens with the 2nd switch rebooting. When the first switch reboots, it doesn't impact the BGP peering on the 2nd one. But when the reverse happens, the 2nd switch is rebooted, all BGP peering on the other switch is also impacted.

I have a maintenance window planned, will do some more detailed investigation first for collecting more information.


Regards,
Sumanta.
Photo of Sumanta Ghosh

Sumanta Ghosh

  • 458 Points 250 badge 2x thumb
Hi Paul

OSPF is not running here, we have direct adjacent i-BGP peering on /30 uplinks. I get this message during reboot, anything to worry about?


Warn:BGP.RIBMgr.ProcStubMsgDataLost

An attempt to process an ATG_I3_INET_ADDR_IND has failed due to local
                                                             resource shortages or indication that the I3 join has lost some
                                                             information for the request
                                                             VR = 0x00000005
                                                             I3 join index = 1
                                                             Data lost flag value = 1
                                                             Local resource ret code = 0


Regards,
Sumanta.