[Systems] Continued packet loss to MIT Media Lab-hosted Sugar Labs web servers may be resolved

[Alex Perez]

Folks,

  or the past ~48 hours, we've had fairly severe, but sporadic, packet 
loss (from 30% to 70%) to the Sugar Labs infrastructure and servers 
(freedom, justice, jita) that host our public-facing websites such as:

* translate.sugarlabs.org
* www.sugarlabs.org
* activities.sugarlabs.org
* all Sugar Labs mailing lists

..as well as the ability to ssh to the servers themselves, plus the 
management BMC/IPMI hosts for the servers. None of the servers 
themselves are actually in a degraded state at the OS level, only the 
network is.

Yesterday afternoon, after rooting around for help on who to contact, I 
opened a support ticket with MIT, and I got a response from them an hour 
or so later, stating: "I'm aware of the issue and am working to 
determine the cause. It is either something failing on the switch or 
something in the optical path. I'm in the process of replacing 
components in the pathway to rule that out."

External monitoring has shown that there have not been any further 
problems with the infrastructure for the past hour and a half, so it may 
be that they have resolved the problem as of this AM. It's just after 
noon where MIT is.

If anybody has any questions, please feel free to e-mail me directly.

Regards,
Alex Perez

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sugarlabs.org/archive/systems/attachments/20190814/407da419/attachment.html>