or Connect
AVS › AVS Forum › Display Devices › CRT Projectors › Curt Palme Down Again?
New Posts  All Forums:Forum Nav:

Curt Palme Down Again? - Page 2

post #31 of 74
Thread Starter 
Quote:
Originally Posted by stefuel View Post

Curt seems to be going down alot lately.

Well at least Sparky is getting a Curt Palme fix

Athanasios
post #32 of 74
Quote:
Originally Posted by nashou66 View Post

Well at least Sparky is getting a Curt Palme fix

Athanasios

Well now, let's see who's banned when the site comes back online...

Just messing with you.
post #33 of 74
Seems to be down now.

Mike
post #34 of 74
Should be OK, I just tried it.
post #35 of 74
He needs to built some redundancy into his site so that when one goes down the other takes over traffic!

Martin
post #36 of 74
We were down for 15 minutes for reboot.

Quote:
Originally Posted by Prehjan View Post

He needs to built some redundancy into his site so that when one goes down the other takes over traffic!

When you build redundancy into a site you have to weigh the cost vs benefit. It costs us nearly nothing to be down for 15 minutes (or even an hour or two for that matter).

To build a truly redundant system means multiple database servers that are sync'ed up real time as a good part of the data is dynamic. Hundreds/Thousands of dollars extra month. Buying dynamic database driven sites that are redundant is a lot of work. I do it enough in my day job. We're not Amazon. There's absolutely no way it's worth it. Very few sites have true redundancy give the cost and complexity to maintain. I'd estimate at least another $5-$10K/year for basic server costs to put even basic redundancy into CurtPalme.com plus all the additional labour of setting it up and maintaining it.

Sorry, not going to happen! If we start selling $50M+/year then possibly but not today!

Kal
post #37 of 74
Quote:
Originally Posted by kal View Post

We were down for 15 minutes for reboot.

When you build redundancy into a site you have to weigh the cost vs benefit. It costs us nearly nothing to be down for 15 minutes (or even an hour or two for that matter).

Kal

Kal that is very true

I was just thinking that he/you can link it with a home connected DSL line just so that it will serve the vital business parts of the site when it goes down!

Anyhow...just food for thought.

Good luck with it

Martin
post #38 of 74
Quote:
Originally Posted by Prehjan View Post

Kal that is very true

I was just thinking that he/you can link it with a home connected DSL line just so that it will serve the vital business parts of the site when it goes down!

Anyhow...just food for thought.

Good luck with it

Martin

But that's my point. There's nothing "vital". People can wait 15 minutes and try again.

What you're saying is that Curt or I should should have dedicate machine running 100% of the time at home on which somehow the dynamic content can be replicated from the live site and kept up to date and that if the main production site goes down that it automatically shifts over to the 'home' site. That's more work than hosting extra redundant servers with the main one.

If it's manual switch over then there's no point: If the server doesn't respond for more than 5 minutes the hosting company reboots it automatically as was done today.

If we're down due to data integrity/corruption problems then the problem would have also been propagated to the backup home server so you're no better off.

EDIT: A home DSL/cable line wouldn't have the bandwidth required either. Never mind the dynamic IP issues. Long story short: By the time you can figure out to get the data pointed somewhere else and get the host tables updated to point traffic correctly, you would have rebooted the main production server hours ago.

Kal
post #39 of 74
Kal

I see and now better understand what you mean.

And yes it has to be done automatically because any other way is a waste of time.

Martin
post #40 of 74
In the end it's balancing act. I could pay 3-4x times as much for a better server or split the web server from the database (two or three different servers to balance load) but as of today, having one reasonably performant dedicated server that needs to be rebooted once every 2 weeks is the best in a cost/performance ratio.

We get about 100K unique visitors/month. A new page is loaded about every 2 seconds, 24 hrs/day 7 days/week. Sometimes we get peaks that drive the server performance into the ground which can (at times) cause performance problems requiring a reboot. That's were we are today. I'm thinking that in 6-12 months I'll move again to a better performing server or split to two servers (one dedicated to the database, one for the app server/HTTP server). We'll see! But for now, a reboot every 2-4 weeks after being down for 15 minutes is fine by me.

Kal
post #41 of 74
Perhaps I should just kill myself. That way not only will I be down, I'll be out as well.
post #42 of 74
I wonder if its physcoidiot giving some payback.
post #43 of 74
Kal, If you want PM me and I might be able to donate some hardware and bandwidth for this project, I run my own datacenter and have a ton of unused hardware that i'm not doing anything with at the moment. Just need to know how much bandwidth you use a day. Thanks Andrew
post #44 of 74
Quote:
Originally Posted by A9A1C1 View Post

Kal, If you want PM me and I might be able to donate some hardware and bandwidth for this project, I run my own datacenter and have a ton of unused hardware that i'm not doing anything with at the moment. Just need to know how much bandwidth you use a day. Thanks Andrew

We use between 300-500Gig/month of bandwidth. Thanks for the offer but any downtimes we have from time to time isn't due to lack of bandwidth or hardware.

Kal
post #45 of 74
I have tried for the last half hour and am not getting on Curt's site is anybody else having problems getting on?
post #46 of 74
Quote:
Originally Posted by bbfarmht View Post

I have tried for the last half hour and am not getting on Curt's site is anybody else having problems getting on?

Yes, the page just tries to load for about 10 minis and then times out for me.
post #47 of 74
From my external monitoring software:

We were down from 2010-05-20 18:24:19 (Eastern Time New York) until 2010-05-20 19:45:33. Approximately 81 minutes.

Welcome to the internet!

Kal
post #48 of 74
Thread Starter 
Curts site wont load up this morning.

Nashou
post #49 of 74
Fixed.

Kal
post #50 of 74
Thread Starter 
That was Fast!

nashou
post #51 of 74
I am down indeed... in California!
post #52 of 74
God. Those border guys can`t stop anybody.
post #53 of 74
Any suggestions on how to get a nec xg751 a to communicate with cpu, Tried to hot wire It with a cable of curts sit,but no luck stumped , I was woundering If nec made a propriatrary cable or can you just buy a standard 9pin-15pin cable to make It work?
post #54 of 74
Before anyone posts here... CurtPalme.com is currently unreachable. Our hosting company's having a major outage at the moment affecting all of their clients (including their own website). They're working on it and hope to have everything back up soon.

Kal
post #55 of 74
Never mind.

wallace
post #56 of 74
Thread Starter 
Kal something is still wrong after the CRT forum fix.

I cant post at all, it says can not post so soon?

Athanasios
post #57 of 74
Thread Starter 
Ok I can post now but its not putting the post at the end of the thread? its on the forum feed back, my post was suppose to be after tri Joel.

Athanasios
post #58 of 74
Thread Starter 
Kal Site is acting weird!!! All threads are loading up blank except one.

Athanasios
post #59 of 74
We are currently experiencing file permission issues which appears to affect all threads *without* attachments (these threads are showing up blank when you click on them). This is currently being investigated by our hosting company - stay tuned!

Kal
post #60 of 74
I've discovered the root cause problem and have identified it to the hosting company.... hopefully it won't take them long to fix it.

Long story short, it has to do with UNIX owner/group permissions of files being set incorrectly for pages that are cached. The resultant cached page is 0 bytes which is turn means that the page comes up blank. Why or who changed this is something I don't know/understand. I did ask for my FTP permissions to be changed slightly last week so that I could create a few new directories outside my 'normal' working area and looking at file dates/logs, it appears to have started back then (feb 13) and for whatever reason isn't really showing up until today. Something else may have changed today.

Hopefully they'll fix it soon. I've given them complete instructions with examples of what's wrong. I'm going to bed!

Kal
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: CRT Projectors
AVS › AVS Forum › Display Devices › CRT Projectors › Curt Palme Down Again?