Astraweb Having Probs..?!? :(us and eu)

[Beck38]

I've caught the us Astraweb 'stalling' a number of times the last couple of days (btw, Giganews keeps rolling at the same time), and this morning about 30min ago, it halted completely. Still zip.

Of course, the first thing I checked was to ping the server, and no problem there (so, the DNS is still alive and correct, and something out there is responding to pings), unless of course, a bad IP has been propagated among the DNS servers (which is possible, but....)

Now, if that were to occur (the DNS server got a bad bit of info), MY local DNS cache wouldn't change unless and until I rebooted... etc.

But nope, the ip address(es) reported by ping are indeed correct.

Well, after some 40 min. or so, they're back up.

Weird that BOTH the us and eu servers would grind to a halt at the same (virtually it appeared) at exactly the same moment.

[heiska]

People have been reporting similar issues all day today on #ab.teevee. I believe it's a DNS problem, but who knows.

Btw congratulations on your 'first' 'post' 'without' 'the' 'annoying' ''.

[cola]

Seems to be all good now.

[Beck38]

BTW, for those not familiar with the San Jose/Santa Clara Country 'infrastructure', they were one of the first municipal governments in the U.S. to build and maintain an industrial fiber loop system throughout their community (back in the early 90's), along with a 'upgraded' city run power grid (long time after the REA projects of the 30's and 40's).

FYI I lived there some 10 years ago.

There have been recent articles lamenting these utilities for failure to keep the system upgraded; as the population grew/exploded (even after the dot.com collapse in 2000) a lot of industrial-type parks intruded deep into residential areas. Plus, the traffic, which is bad to begin with in the SF bay area, but Santa Clara county passed on BART rapid transit in the 1970's, so everything is cars cars cars.

In short, lots of the larger customers have been moving their IT operations 'down the road' a few miles to get out of the clutter (and telecom/electric grid has improved quite a bit over 20 years).

[Rogi]

im having issues with astraweb aswell right now. Really slow and often dropping connection whenever an artical finished downloading. No clue whats going on.

This is EU-ssl.

Ohwell guess ill be using my isps shitty 7 day retetion news server untill this is fixed.

[w1zardboy]

other than the problem you are having now, has Astraweb been ok for you?

I've only just signed up with them 2 days ago and am using it instead of my giganews account, just until i see how it performs, with the view of cancelling my giganews account. $30 for me was becoming too much to justify.

I'm not rushing into cancelling it for a few days tho, as my supersearch newsleecher is tied to the giganews account and dont want to make any rash decisions.

[Beck38]

What is (somewhat) fascinating to me is the amount of 'fills' I get from the Astra EU, with Astra US as my primary.

Those two ought to be primary peers to each other, anything 'missing' on one should be backfilled by the other. And I'm not talking about something posted a couple days ago, it's something in the 400-600 days back.

The US server plant seems to suffer from what I call 'fades', where missing articles will occur, generally across the entire newsgroups, for several minutes to hours. At the same time, the EU server doesn't.

Now, if they were properly peering off each other, it would 'backfill'. But it doesn't. Ever. This might have something to do with message ID's (the long long long alphanumeric 'serial number' that's stamped with each usenet message upon creation).

When usenet was created, the size of that number (if memory serves) 28 digits long. That was fine for about 25 years. Not any more. There are techniques to 're-use' the numbers, but that only holds a little breathing room.

Just like IP addresses, this is something that will 'eventually' get 'fixed'. As is typical (with humans), it will rum right up against a crises before permanent action is taken.

[w1zardboy]

thanks beck for the different slant. intersting tidbit on the numbers thing :)

I've been using the eu server the past couple of days, and havent found any difference (in relation to incomplete files) in what i can get from there, as opposed to what i can get from giganews.

To be honest it was the growing number of incompletes i was coming up against, which made me reconsider moving ( i was very giganews loyal from the outset, i still cant see why though :frusty: )

but now the move has been made i think i'll give it a week or so, just to settle into it. Have never really thought of having the US in standby to make any fills though. but as you say , its been the other way round for you, anyhow.... great tip!!

[Beck38]

What's interesting (from my standpoint), is that although I'm about 500 'linear' miles from the US Astra, and close to 7000 miles from the EU/Amsterdam server, 'internet wise' the number of 'hops' (by tracert) and delay (round-trip pings) is almost reversed.

I.E., although US Astra pings are VERY quick millisecond wise (makes sense), the number of hops is a bit excessive (around 14). But I do have a Global Crossing port just down the road from me, and that results in a quick path to Amsterdam (11 hops) but the delay is sluggish (about triple that of the US Astra).

Upon occasion, I have run for days off the EU server, 'just fer fun'.