Astraweb (US and EU) SSL Connection Problems

[Beck38]

This weeks 'fall down go boom' problem with Astraweb appears to be SSL (both port 563 and 443) connection errors (refused connection).

I've actually got an SSL/563 connection operating (10 connections out of the 20 allowed), having started it about some 6 hours ago, but when I tried starting some additional ones I got a connection refusal. I do this all the time, sometimes starting up some flows from, say the EU server or some adds from the US one.

A non-SSL (port 119) connections works just fine.

Opened a ticket at Astra.

[Beck38]

Actually got a fairly quick response from Astra, and indeed, sometime over the last 8 hours or so it did get fixed (on both server plants).

No idea on exactly what was 'responsible', but as Astra uses 'GoDaddy' for the SSL connections, it was probably something in the SSL certificate handshaking or something in that area that caused the problem.

Interestingly, the 10 SSL connections I had up and running (on Astra/US) 'before' this failure occurred, kept running throughout the entire failure. Okie Dokie.

[pinhead1000]

i cant connect at all to the servers been trying all day and nothing on the ssl or normal

[Beck38]

i cant connect at all to the servers been trying all day and nothing on the ssl or normal

No problem on either US or EU, SSL or otherwise, over the last 4+ hours, with long-duration connections or short (posting or leeching).

I'd look at some other potential problems with your set-up or ISP.

[Hypatia]

http://www.reddit.com/r/usenet/comme...nnecting_with/

[piercerseth]

The routes are messed up. My packets don't normally transit Level3's pipes, now they are and getting lost at the edge. EU servers are fine as previously mentioned.


EDIT: Seems to be working again.

[Beck38]

The routes are messed up. My packets don't normally transit Level3's pipes, now they are and getting lost at the edge. EU servers are fine as previously mentioned.

With me (on the west coast of US to Astra/US/SanJose) 99% of the time transit L3, only a couple times a month I'll see SprintNet, but then again this is through a VPN service that has an 'out-port' in San Francisco.

Sound to me like some carrier out there made a 'slight' change to their routing tables that threw the proverbial monkey wrench into the works. A few months back it was L3 that ran into having older Cisco routers that met their limit on routing table ram that caused nothing short of massive problems around SF/SJ. It got some actual press for a few days, they had a 'quick work around' that sent traffic from SF to LA then back again for a few days and then did some fix that pub things 'back to normal'.

If folks with the current problem can shift (if they have it) their VPN service to another out-port, they may be able to work around it (if it lasts into days instead of hours). Then again, the carriers might actually figure out they have (or generated) the problem and maybe figure out the 'fix' or a work-around. Only a bit of time will tell at this point.

[rabcfc]

Very up and down for me, when it works its awesome but doesn't happen often, using ssl

[Beck38]

There was some 'talk' on 'Broadband Review' (DSL Reports) but it all seems to have gone away (as in, it all appears to be working again). Sounds like it may have been a short 'glitch' in some router somewhere, that's been kicked back into line.

Never saw anything go out of whack here, was doing my usual 24/7 uploading routine (with occasional testing and such). I'm taking a break right at the moment as I'm buying a new house and I've really got to have my 'wits' about me, at least for the next few days.

[demod8698]

I had issues as well, but EU is working fine. Can anyone verify without a doubt US is back up? I just tried and no dice.