August Backbone Engineering Report
Jordan,
I notice that there was no mention in your report of the serious outage NEARnet experienced early in the month. Do such outages get reported via some other mechanism?
- John
John, We generally do not report specific outages in the monthly engineering report, unless they are causes or effects of changes in network software/hardware. With respect to the outage you describe, there were two routing problems which surfaced at the beginning of the month that caused some outages. We inadvertantly neglected to include these. Thanks for pointing this out. To clarify: a bug in the EGP protocol code of our T3 rcp_routed software induced a serious outage at both NEARnet and SESQUInet on 8/11, and resulted in at least one change to the T3-rcp_routed software. This occured at about the same time that we exceeded the 8KB EGP update size in outbound network announcements, which caused problems with a few regional routers, and complicated the diagnosis of the T3 routing software bug. As an aside, we have been maintaining a current set of release notes on the T3 network routing software that describes the bug fixes and changes in each release. This is available via anonymous on ftp.ans.net in the file /pub/info/T3-rcp_routed/Release-Notes. We hope that people find this information useful and welcome feedback from the community. Finally, our practice for outage reporting has been to post real time T3/T1 outage reports involving regional network and service provider attachments to the NSR mailing list and provide additional information on the regional-techs and nwg mailing lists. Jordan
Jordan, Thanks for your note. I'd like to discuss how NEARnet and ANS could improve our communication the next time we meet. I think regular reports, at the level of engineering and operations, are a good means of exchanging information between organizations and between the technical and managerial staff of those organizations. - John
participants (2)
-
Jordan Becker
-
jrugo@nic.near.net