Packet loss at gin NTT

[Amileah]

I work for an ISP so good traces are exactly what we need to identify issues on the network. And sending the email to NTT was smart. He could also sent in those to Mediacom as well. Its way pass the egress point in Minneapolis but sometime ISPs have agreements with other carriers and are willing to look into issues if they are a partner.

Mediacom shrugged it off as not their problem because it was outside their network by the time packet loss occurred. That is kind of what I figured from them though. I have the business support info and I could talk to them but I don't want to clog up business lines for a viddiya game especially when its my home having issues and not my business.

Update --NTT-GIN did also see packet loss, they are actively monitoring it.

[Annabellechu]

More packet loss from snjsca04 today and I got some yesterday. What's the best route to contact them?

[Psythe]

Yeah this afternoon the snjsca04 hop is losing around 12% of its packets. For almost 3 hours. This morning around 2AM pacific time it was averaging 9% loss for over an hour. The last hop before the DC has been crazy with 40 - 65% loss whatever device that is(load balancer, gateway, whatever). There is at least 4 threads with network tests across the first couple of pages. All of them showing that SE should be all over their DC and/or Network providers behinds because there is no way that is within acceptable SLA. SE already has an issue with things they can not do much about but this is what that could alleviate a large group of issues with connectivity that is not related to the Queue service.

I ran tests from servers all over the globe (6 locations). I have access servers across multiple DC's and along with the various 3rd party remote testing tools all of them have shown the same issue.

[technole]

I've been getting logged off more than twice today. Did the traceroute and it's definitely NTT's SNJSCA04 being a bugger. NTT hosts the NA data centers in Sacramento, so it's really in their hands.

[Annabellechu]

I've been getting logged off more than twice today. Did the traceroute and it's definitely NTT's SNJSCA04 being a bugger. NTT hosts the NA data centers in Sacramento, so it's really in their hands.

I've had those 2 mostly so it includes the same one as you:
ae-2.r23.chcgil09.us.bb.gin.ntt.net
ae-1.r24.snjsca04.us.bb.gin.ntt.net

[technole]

I've had those 2 mostly so it includes the same one as you:
ae-2.r23.chcgil09.us.bb.gin.ntt.net
ae-1.r24.snjsca04.us.bb.gin.ntt.net

I posted a thread about this a few days ago https://forum.square-enix.com/ffxiv/...-s-own-network

I ran traces from various locations using a VPN, the only hop that was consistent and also consistently bad was 204.2.229.230.
Starting from different locations sends you through various NTT.net servers, while seeing loss on their servers was common, it was about 90% of the traces on hitting 204.2.229.230
So that might be an issue with NTT.net's connection to 204.2.229.230. As it appears NTT.net is the internet provider for the lobby server, hopefully SE is already working with them to get the issue fixed

I contacted NTT NoC and they got back to me and was very kind. Traced through Dallas hub to snjsca04 and didn't see anymore packet loss on their network so it seemed like that's smoothed out now, they also believe the SE lobby server could be throttling causing the loss % at that final hop IP you mention. Definitely was much smoother this evening. Hopefully the maintenance on the lobby server tonight will help.

[Devourn]

What the? Do they use duct tape and bubblegum to hold their network together? They constantly have problems. eyeroll.