Intermittent Outage?

Geoff Winnington-Ball (RIP) · #1 26-08-07, 16:59

EVERYBODY --

This weekend I've been receiving several reports, mostly from Australia but here and there elsewhere as well, that some have been getting Server not found or similar errors when they have been trying to access http://www.mapleleafup.org/ and other MLU URLs such as this Forum.

One chap indicated that he went through his bookmarks at the time and nothing else seemed to be affected, just us; then, when he tried again, it worked.

If this is the case, it's probably a DNS routing problem somewhere in the server farm operated by my hosting service - but it may also be between them and the main internet trunking. It's going to be tough to nail down!

Two things:

1. When this happens, could you email me at sunray@mapleleafup.org and send me the message you're receiving, noting your local time? That will help.

2. Having dealt with DNS problems before, try typing in the address without the "www.", ie http://mapleleafup.org/forums; if that works, use your link to go back to the normal one and recheck that. In any case, send me the results of that little test as in item (1), please - every little bit helps!

Thanks for this - I'll try and get to the bottom of it as quickly as possible.

Jif

Keith Webb · #2 26-08-07, 21:29

Or as you Canucks say 'OAT-age"

Yes, I've experienced this - just over a year ago it was quite severe, then no problems and then a few days ago it started again - the MLU address doesn't connect most of the time which is of course very frustrating.

Because some of us can't get through a lot of the time it may take a while for people to even see this to report it.

Anyway it goes to show how important communications between this community are!

I have a theory it may be something to do with servers experiencing massive spam attacks where for periods of time they block certain IP addresses - Oldcmp.net was until last weekend located on a server in the US but we have had to move it because of this problem which resulted in our site being constantly infected and even having email trouble.

Hopefully the hosting people are able to sort it out.

Plushy · #3 27-08-07, 02:30

I have been experiancing this problem as well but my local Isp was having other problems as well so i thought this was just an after affect .

Fingers crossed it gets sorted out soon .

cletrac (RIP) · #4 27-08-07, 03:08

I've been having that trouble for several days now. This is the message window that pops up. I had to click on my MLU bookmark 11 times before it connected just now.

#5 27-08-07, 05:03

I've had message pages pop up too ... not able to access etc. Most recent was a couple of hours ago but I refreshed the screen and MLU came back but I did completely losethe PM I had just hit send on and a can't access page popped up. That PM did not go through.

I've noticed that perhaps the last couple of days or so that sometimes the pages are extremely slow to load again (both main page forum as well as page to page linking)

Hopefully bugs will get figured out and fixed soon.

David Ellery · #6 27-08-07, 06:41

Thank goodness it wasn't just happening here,
I thought maybe NZ had drifted a bit more South and snapped the cable.

Phew, started to panic.

cliff · #7 27-08-07, 08:17

I have not been having these probs but one day last week it would not open threads even though I was on site and could get into the different sections like softskins.

#8 27-08-07, 08:50

Maybe when we find the culprit we can all

it?

Keith Webb · #9 27-08-07, 09:50

But has gone dead again since.

cliff · #10 27-08-07, 10:55

fine for me just now Keith

Keith Webb · #11 27-08-07, 11:45

Just started working again too

Geoff Winnington-Ball (RIP) · #12 27-08-07, 11:57

OK right now for me and no error messages this weekend at all HERE... but that doesn't mean there isn't a bug.

From what I've been able to learn over the years of doing this, Assholes International® has made a competitive sport out of bombing whole sections of the internet to bring them down. Flooding individual servers with computer-generated requests can lock them up altogether, and MLU HAS been a victim of that before; at that point there's no choice but to shut it all down for a while until the attacks shift focus.

MLU is hosted on a huge, shared server - it's the only way to keep costs down to manageable levels. Sure, I could go to a dedicated server to escape being caught by attacks on other sites on this server, but the cost would be five to ten times what it is now - obviously a non-starter.

Even if the Lotto Faerie paid me a visit in the night and I could arrange this, there's no telling that the morons wouldn't, as they have in the past, go after individual routers in the network, or that flood attacks on other servers wouldn't lock up those routers. The latter are like merges on the highway - we might be a major artery, but sooner or later we're going to have to merge in with a larger highway to funnel traffic in or out everywhere else... otherwise we'd be an intranet instead of being on the internet.

I'll keep y'all posted...

#13 27-08-07, 13:32

I wonder sometimes if this stuff has anything to do with this being a 7 year old version of the program and if maybe the upgrade would solve a lot of the problems. I realize it costs money to upgrade and I have none to shoot your way for that, but if I did, I would. Wish I could.

Geoff Winnington-Ball (RIP) · #14 27-08-07, 13:55

Quote:

Originally posted by Vets Dottir 2nd
I wonder sometimes if this stuff has anything to do with this being a 7 year old version of the program and if maybe the upgrade would solve a lot of the problems. I realize it costs money to upgrade and I have none to shoot your way for that, but if I did, I would. Wish I could.

No, nameserver troubles cannot be caused by software running within the domain.

Yes, we need an upgrade. V2.3.5 came out on '04, but V3.x has been out long enough for the bugs to have been sprayed, so one of these days... it might at least speed up the forum a bit, but that's also a function of both the amount of daily traffic and the huge database of stored posts. (Disabling the search function WOULD speed things up a bit, but I don't particularly like being burnt at the stake, so I'll pass on that for now.)

servicepub (RIP) · #15 27-08-07, 15:00

I had problems all weekend - only successfully entered the forum once. I thought that this was a repeat of a similar problem that I was experiencing last year and, as it was always on weekends, I put it down to something wrong with my home PC.

Hendrik van Oorspronk · #16 27-08-07, 22:16

I had the same problems, but I had a link from Hanno when I had the same problem before, it still worked.

Green Greetings Hendrik

#17 27-08-07, 22:32

Quote:

Originally posted by Geoff Winnington-Ball
No, nameserver troubles cannot be caused by software running within the domain.

Yes, we need an upgrade. V2.3.5 came out on '04, but V3.x has been out long enough for the bugs to have been sprayed, so one of these days... it might at least speed up the forum a bit, but that's also a function of both the amount of daily traffic and the huge database of stored posts. (Disabling the search function WOULD speed things up a bit, but I don't particularly like being burnt at the stake, so I'll pass on that for now.)

This is all beyond me wee brain to follow and understand

I just wondered if it was the software a problem instead of the server problem ... but you say it can't be so I'll be quiet now

Grant Bowker · #18 27-08-07, 22:34

So far no problems either at home or work for MLU access, but it does explain why there has been a lull in postings. I had assumed that everyone was working furiously on restorations..

#19 27-08-07, 22:37

PS Sunray ... I surely would not want you burnt at the stake .... it would deny me the use of my

on you ...

Geoff Winnington-Ball (RIP) · #20 27-08-07, 23:55

Bloody hell, I've been at this since 1715Z, and it's now 2255Z... but I think I've found the glitch.

(1) 1715Z - I refreshed and got a 502 Bad Gateway error. Refreshing brought me a 504 Gateway Timeout error. I experimented with different browsers and different machines, same error(s).

(2) 1800Z - I opened a formal trouble ticket with my hosting service, detailing responses etc.

(3) 1845Z - Host writes back to say that this error(s) is so rare they had to look it up, and suggested it was at my (local) end. I made rude comments to myself. They added it might be a good idea to run a traceroute on these services to see where they ran parallel to one another or diverted. I DID have a back door into the Forum, which was a place to start.

(4) 1900Z to 2030Z - I one-upped the buggers... that'll teach 'em to tap-dance with an Honourary Gunner! Went into WHOIS lookup AND did a traceroute on all services, whereupon I found that half had reverted back to an old, disconnected server but still worked (haven't figured that one out yet, it's the GPO ACK's job and he's drunk again) and the other half still with proper DNS routing. Yet going off in all directions.

(5) 2035Z - printouts in a text file, I mailed back the hosting dweebs and said "Sort THIS, you..." (I was polite, didn't want them turning me off altogether).

(6) 2230Z - They reverts ALL DNS pointers back to what they should be on all 16 domains I host. Tested and functioning, but not, repeat not verified yet.

(7) 2245Z - Another cup of tea. and commence report.

ADDENDUM: Please advise me ASAP if you experience any more like troubles so I can trace them accordingly, now that I've wasted a whole bloody afternoon learning far more about DNS Devils® than I really wanted to know.

DNS Jif

#21 28-08-07, 00:02

Amazing how tedious and hard it is to try and get some answers and corrections. Good job G

cletrac (RIP) · #22 28-08-07, 00:56

I just had to try ten times before it hooked up.

Geoff Winnington-Ball (RIP) · #23 28-08-07, 01:55

Based on your comment, David, I just reran a set of tests of 10 domains; all end up where they should but are still routed in the process through a dysfunctional server. This could be causing the delay you saw.

I've reopened the ticket with my hosts, copied in a text file showing the DNS routing and asked them to dig deeper. They did SOMETHING last time but not enough.

Starting to feel really old right now, am going out for a walk. Hold the fort, everyone. I'll stay on this until I have an answer, but I've already been sitting in front of this screen for 14 hours.

Jif

cletrac (RIP) · #24 28-08-07, 02:01

It's improving. It just hooked up on the second try.

Pedr · #25 28-08-07, 03:54

G'day Boss

Haven't had any trouble here for me for a long time ( touches wooden objects profussely )

Am on 100Mbps Broadband connection with forum index page as home page inside sign in. Works fine.....but I'm on Firefox and not IE, as IE has always been a problem for me.

Typical of host...."Not here, must be at your end"..in other words...can't be f@$ked looking, go away it might sort itself out"

Fingers crossed that you can find the glitch and get it sorted without too much more grief.

Pedr

PS have worked out why smilies don't animate everytime. Firewall settings.

Geoff Winnington-Ball (RIP) · #26 28-08-07, 04:52

Nothing back from them yet. Moot point, going-to-ground until oh-dark-thirty.

SUNRAY OFF EFF 0400Z

#27 28-08-07, 06:03

Quote:

Originally posted by Geoff Winnington-Ball
Nothing back from them yet. Moot point, going-to-ground until oh-dark-thirty.

SUNRAY OFF EFF 0400Z

Good plan ... your computer marathon of all day at work and constantly at home working online the last few days calls for a good break and sleep. MLU is glitchy at the moment but still accessible a lot while its being sorted out. Its not like its completely down at the moment.

Sleep well ... your health number one!!! ...

cliff · #28 28-08-07, 07:35

Quote:

Originally posted by Geoff Winnington-Ball

Starting to feel really old right now, am going out for a walk. Hold the fort, everyone. I'll stay on this until I have an answer, but I've already been sitting in front of this screen for 14 hours.

Jif

LOl but Jif I tort you vas allreddy old for long time!

Keith Webb · #29 28-08-07, 11:23

It's been down again.

Geoff Winnington-Ball (RIP) · #30 28-08-07, 11:38

The following received during the night:

Quote:

Actually we don't see any traces of ns1/ns2.dnsrouter.com left in the zone files. What is possible is that the DNS zones are cached where you're running them through, in which case it depends on when their cache expires (usually within a few hours).

The good sign is that the host has totally corrected the original problem at their end, but the bad news is that it may be hanging in one of the relay routers in the system due to erroneous information having been dumped on the system by the initial 'relapse'. If so, it may clear itself today, but if not, I may have to force a change in the NS IPs at the source (no, don't ask me to explain).

I'll keep an eyeball on it today and YOU keep feeding me trouble reports - just ensure you have flushed YOUR cache before you draw up a fresh page.

Jif