Internet routing issue

[S6PNJ]

I've been having a strange internet issue for about the last 1-2 weeks. I am currently unable to synchronise some email accounts or access the webmail URL. This is not all of my email accounts, just those associated with a specific domain, which I understand has recently been migrated to a new service provider.

Gigaclear is my ISP and I'm also able to get onto another Gigaclear instance a few hundred meters away (local Village Hall), so in theory, apart from locally terminating in a possible different street cabinet, they amalgamate onto common network infrastructure pretty soon I'd imagine.

My home/laptop setup is Win 10 home, MS Office 2021 and Google Chrome - all fully up to date with patches / software updates etc. Same kit used in both locations.

If I go to my village hall, my laptop can synch with the emails and can also get onto the webmail URL and successfully login. I can't do either from my home location, ie I can't access the URL and I get an MS Outlook synch error, rather than it reaching both access points and failing authorisation.

If I tracert from the Village Hall, I get:

1 2 ms 2 ms 2 ms router.router.local [192.168.1.254]
2 12 ms 9 ms 8 ms 46-227-148-254.customer.gigaclear.net [46.227.148.254]
3 7 ms 10 ms 8 ms 10.104.0.149
4 9 ms 9 ms 15 ms 10.104.0.148
5 14 ms 9 ms 9 ms lonap-gw1.sotaconnect.net [5.57.81.210]
6 23 ms 16 ms 14 ms xe-0-1-0.rt1.ksp.racksrv.com [83.137.224.126]
7 20 ms 12 ms 9 ms cp165175.hpdns.net [91.238.165.175]
trace complete



If I tracert from my home, I get:

1 5 ms 3 ms 6 ms Gigaclear00528 [192.168.0.1]
2 25 ms 9 ms 10 ms 46-227-148-254.customer.gigaclear.net [46.227.148.254]
3 9 ms 6 ms 9 ms 10.104.0.149
4 8 ms 7 ms 7 ms 10.104.0.148
5 9 ms 8 ms 8 ms 195.66.227.142
6 11 ms 11 ms 14 ms xe-0-1-1.rt1.ksp.racksrv.com [83.137.224.122]
7 * * * Request timed out.
and it then continues to time out until I Ctrl C to get out.


I'm assuming the 10. addresses are Gigaclear infrastructure but the next 5. (Village Hall) and 195. (home) are not on Gigaclear's network? Is this likely or not likely to be the case?

In fault-finding / troubleshooting Gigaclear has asked me to remove power to my internet kit for 15 mins (I did 30 min on Mon eve) and to power up my fibre before my router. Gigaclear are currently suggesting it is an email issue (I don't think it can be as I can synch from other locations and other users of those email accounts are not having issues) or possibly a router issue and for me to contact Linksys (it's a Gigaclear supplied router, so in my eyes, still a Gigaclear issue) and progress with them.

From what I can see it is far outside of my home network but somehow associated with my 'setup' via Gigaclear.

Any techy types out there who have seen similar or have any ideas of what might be causing the issue?

[ffc]

Looking at the traceroutes the issue is at rackspace.

Weirdly the rackspace router can reach the destination device via one interface (xe-0-1-0.rt1.ksp.racksrv.com [83.137.224.126]) but not the other (xe-0-1-1.rt1.ksp.racksrv.com [83.137.224.122]).

I guess you can try calling Rackspace but this isn't a problem with Gigaclear.

[S6PNJ]

Thanks for the steer.

I was thinking the issue was here:

5 14 ms 9 ms 9 ms lonap-gw1.sotaconnect.net [5.57.81.210]
vs
5 9 ms 8 ms 8 ms 195.66.227.142

so prior to the Rackspace addresses. Are they also Rackspace locations? or does the fact that both routes get to Rackspace means the issue is from there onwards?

Is this - https://www.rackspace.com/en-gb the Rackspace I should be contacting or is there another?

It's all still a mystery to me - hence many questions.

[outnumbered]

I think you'd have to start with Gigaclear to make progress with this, and show them the evidence you've found. They're more likely to be able to fix it and/or contact the upstream site successfully I would think.

Another interesting factor might be your external IP address in both locations. What does http://ip4.me/ tell you ? (and despite the trace showing a problem with IPv4 connectivity are you sure there's no IPv6 involved ?)

My best guess would be that there's been abuse spotted from one of the source IPs in the past, and it's on some blacklist at the mailhost.

[sunbeam alpine]

ffc said:

Looking at the traceroutes the issue is at rackspace.

Weirdly the rackspace router can reach the destination device via one interface (xe-0-1-0.rt1.ksp.racksrv.com [83.137.224.126]) but not the other (xe-0-1-1.rt1.ksp.racksrv.com [83.137.224.122]).

I guess you can try calling Rackspace but this isn't a problem with Gigaclear.

If the OP can supply the tracert results to Gigaclear tech support, shouldn't Gigaclear be raising this with Rackspace?

[S6PNJ]

Thanks, I've shown / sent Gigaclear the tracert info and they are (currently) basically saying it's my issue (ie email or Linksys router) and not their issue.

I'm 99% sure it's not an email issue as I can access it from elsewhere
If a Linksys router issue, then that was given to me by Gigaclear so again, their issue from my POV.

I'm pushing it with Gigaclear tech and am awaiting a call / email back from them following another call to them yesterday.

The ip4me tells me my WAN address (it's a static IP so I'm not posting it here!) and tells me I'm connected IPv4. If I do the IPv6 test, it fails - site can't be reached. I'm too busy to currently 'nip out' to the village hall to check from there but can try later - I suspect I'll get a v4 only from there as theirs is an older router than what I'm using here (and is also not Linksys)

And yes, I firmly think Gigaclear should be pushing this, not me, but as I'm the one affected and am but a small fish in Gigaclear's pond, it's pretty much up to me to progress and push as hard / far as I can.

[Scarletpimpofnel]

ffc said:

Looking at the traceroutes the issue is at rackspace.

Weirdly the rackspace router can reach the destination device via one interface (xe-0-1-0.rt1.ksp.racksrv.com [83.137.224.126]) but not the other (xe-0-1-1.rt1.ksp.racksrv.com [83.137.224.122]).

I guess you can try calling Rackspace but this isn't a problem with Gigaclear.

Years ago I had servers with 2+ NICs, each to access certain networks. The server had to have it's routing table modify to add relevant entries to correctly route traffic to the right NIC to get to the right destination. It caused all manner of intermittent issues if such routes weren't set up. I think rackspace have this issue, traffic going to the wrong NIC then going nowhere. With no routing table it can sometimes work and sometimes not.

[theboss]

The host with the routing issue is racksrv.com not rackspace; the fact I can only load bits of their website doesn't instil much confidence.

Neither does companies house for that matter - they are in liquidation

https://find-and-update.company-information.servic...

[S6PNJ]

Ah! the plot thickens!!

If I type racksrv.com into my browser, I get hostpresto.com - guess who the new host provider my email chap has just gone over to! Yup, you guessed it! I'll try contacting them but I don't hold out much hope...

[theboss]

I'd be taking an offline copy of all my emails if I were you!

[Mr Pointy]

theboss said:

I'd be taking an offline copy of all my emails if I were you!

And investigating MS365 & Gmail.

[S6PNJ]

Mr Pointy said:

theboss said:

I'd be taking an offline copy of all my emails if I were you!

And investigating MS365 & Gmail.

It's not my email or website domain, I am but a user of that domain.

[S6PNJ]

theboss said:

The host with the routing issue is racksrv.com not rackspace; the fact I can only load bits of their website doesn't instil much confidence.

Neither does companies house for that matter - they are in liquidation

https://find-and-update.company-information.servic...

As I say, resolves to Hostpresto.com for me which says 'Enix Ltd trading as...' - https://find-and-update.company-information.servic... so maybe a phoenix?

Just done another search / test:

racksvr.com resolves to hostpresto.com but racksrv.com/index.php resolves to the Enix Ltd / Racksrv webpage that only slightly loads as mentioned.

[theboss]

S6PNJ said:

As I say, resolves to Hostpresto.com for me which says 'Enix Ltd trading as...' - https://find-and-update.company-information.servic... so maybe a phoenix?

Just done another search / test:

racksvr.com resolves to hostpresto.com but racksrv.com/index.php resolves to the Enix Ltd / Racksrv webpage that only slightly loads as mentioned.

Seems like a probable acquisition, in any case somebody has kept the kit running throughout a liquidation process whereas a colo would probably just turn the rack off once a certain level of invoices were in default!

That's probably where you'd want to direct your query in that case.

I've seen issues like this quite a few times before, not just with small ISPs but with the biggest carriers who aren't interested in users trying to help draw their attention to network anomalies. Hopefully a smaller one is more receptive.

Any attempt to engage an ISP usually goes nowhere as above, they'll just say its off network and nothing they can do. No criticism of Gigaclear, again I've seen this at all service levels.