Reminds me on old days when acls didn't have numbered lines During the migrations we had to maintain production and temp acl and change the distribute statements to use temp acl before deleting the production acl and pasting an updated version of it than put the production one back in and update the temp acl Nightmare and very error prone
adam -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Nick Hilliard Sent: Wednesday, February 22, 2012 11:12 AM To: [email protected] Subject: Re: [c-nsp] automatic bgp route refresh On 22/02/2012 09:11, Phil Mayers wrote: > Even more so, it can be disastrous, causing traffic blackholing, if you're > editing communities or attributes required to "make things go" and need to > do this in a multi-step way e.g. > > route map blah 10 > ! There's no "no set community 1:2", thanks IOS > no set community > ! oops a route refresh happens now I've sent all my routes > ! with no community, and the community is necessary in this case > > ! a few seconds later... > set community 1:3 sigh, yes. This is just one fine example among many of why atomic (or even pseudo-atomic) edit / commit / revert is a necessary tool on routers. "reload in 5" still sucks as a means of working around this. Nick _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/ _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
