Nightly freeze?

It happened again at 0906pm MDT.
1a.gif
 
flyingcheesehead said:
Went down at 9:32 CDT tonight, still down at 9:37, back up at 9:41.
Click to expand...
Right about the same time for me with the computer clock showing 10:31 EDT. I just gave up and went off to take out my frustration by abusing some packing boxes. :D

It's time to address this or move. What other options are there?
 
It is not just a nightly problem. I have been noticing the same type of delay at 9:30am CDT as well. I am 3illing to be that there is someone on one of the virtual servers that has set up a chron job to occur every 12 hours and what we are seeing is server overload (Jesse has confirmed this for the evening).

I figure we have two choices. Move to a real server or just suck it up. But it is damn annoying, I would hope that we could move to a better server.

Chuck any update or is this not even really on your radar screen?
 
I've been looking at options. However, the only option better than a virtual server is a dedicated server, and dedicated servers cost more.

I've been looking into matrixing - a grid based system that would be cheaper, hosted vs. semi-dedicated, but supposedly would be a better overall hosting system. However, my initial research indicates that a lot of that "better" is more hype than reality.

Cheaper would be better. Anything more expensive is going to fail the "can we live with it as it is now" test, I think.
 
If we stick with our current host, our cost would go up from approximately $50.00 a month to $80.00 a month, a 62% increase.

Right now, we have funds in the bank sufficient to operate the forum for another six months before we need to do more fund raising (thats a mix of MC contributions and general member contributions). Moving to a dedicated server would drop that to 3 months.

One option I'm considering is setting up a matrix account at my own expense, moving PoA there on a trial run to see how it goes, but keeping the current account active.

My main concern is how well this new setup would work with mySql, and by their own admission, mysql doesn't do so hot on their setup, simply because of the architecture involved. When it comes down to it, mysql is the heart and soul of vBulletin, so if it didn't run well on the grid, there would be nothing I could do to help it.

Going dedicated, aside from the cost concerns, brings others. I am NOT a linux expert, and we would need to recruit one whom we could trust absolutely to "hold the keys" and make sure we were doing everything right. I've already made a bit of a mess of the current one... but it limps along for now.

So cost, expertise - both are major concerns to be addressed before we make any major changes.

Oh - yeah - and the fact that everyone who runs this is doing so voluntarily, and we have lives, jobs, and expensive hobbies like flying... ;)
 
Greebo said:
Going dedicated, aside from the cost concerns, brings others. I am NOT a linux expert, and we would need to recruit one whom we could trust absolutely to "hold the keys" and make sure we were doing everything right. I've already made a bit of a mess of the current one... but it limps along for now.
Click to expand...

There are several people on this forum with the Linux skills to take care of this problem. I don't think moving to a dedicated server and having it properly maintained would be an issue at all.

All of us have jobs. All of us have lifes. That said this performance problem does not take much time to resolve at all. The community just needs the ability to solve the problem.

I do not think that just one individual should hold the keys. There are plenty of ways to set things up to where this wouldn't be the case. Personally I think a "technical board" per say should be setup where some trusted long term members can respond to these problems. Things will be properly maintained and not one person can pack up and run.
 
jesse said:
There are several people on this forum with the Linux skills to take care of this problem. I don't think moving to a dedicated server and having it properly maintained would be an issue at all.

All of us have jobs. All of us have lifes. That said this performance problem does not take much time to resolve at all. The community just needs the ability to solve the problem.

I do not think that just one individual should hold the keys. There are plenty of ways to set things up to where this wouldn't be the case. Personally I think a "technical board" per say should be setup where some trusted long term members can respond to these problems. Things will be properly maintained and not one person can pack up and run.
Click to expand...
Regarding 1 person with all the keys, I actually agree. Adam Z is, as time allows, working on establishing PoA as a non-profit (non tax exempt) entity.

Once that is done, I'll be a lot more comfortable making copies of the keys. Right now, however, since everything remains in my name, I'm a little paranoid.

However, the technical matters are secondary. Cost is the primary issue.
 
Problem has been coming and going for the last two hours. The double posts and double threads right now are the result.
 
how annoying. any chance we could move the chat to a different server (perhaps hosted by an established member) so that we could enjoy uninterrupted conversation?
 
jesse said:
Problem has been coming and going for the last two hours. The double posts and double threads right now are the result.
Click to expand...
Yep, it just happen to me. And, I'm on a slow air card in the car. It took over five minutes for me to get a return on new posts yet it managed to post my new thread twice.
 
FYI:

I'm looking in to setting up a test run on another host. Its in the beginning stages right now, but I just wanted you all to know that we're not ignoring this.

Chuck
 
Greebo said:
FYI:

I'm looking in to setting up a test run on another host. Its in the beginning stages right now, but I just wanted you all to know that we're not ignoring this.

Chuck
Click to expand...
That remotely sounds like a campaign promise. :)
 
Vote for me, because if you don't, it doesn't matter, cause I'm still one of the ****ards in charge!
 
I noticed two today. Both lasted about 5 minutes. First was this morning at 8:40am and the second was at 7:46pm CDT on both. I have noticed morning period ones. Usually also around 9 to 9:30am CDT, 12 hours after the nightly freeze.
 
Right now I'm experimenting with different ways we could move to a test run on a new server.

The way I prefer - a clean code install and a data migration - will be a very slow process - but might be worth it. However, it would mean that some of the conveniences we're used to here would go away, for a while.

The alternative is to just copy the code in its current state and do a normal upgrade, but I'm not inclined to do that. New server, clean code. Has a lot of appeal.

Testing is just beginning. I'll be keeping everyone appraised as developments develop. ;)

Maybe in a week or so we'll be ready to load test the new platform.
 
Greebo said:
Right now I'm experimenting with different ways we could move to a test run on a new server.

The way I prefer - a clean code install and a data migration - will be a very slow process - but might be worth it. However, it would mean that some of the conveniences we're used to here would go away, for a while.
Click to expand...
Please elaborate!
 
I will when its not so late.

Tho right now I think perhaps the *first* test should be with the code exactly as it is now, and then we can either prove that its something on our end, or it isn't...
 
Greebo said:
I will when its not so late.

Tho right now I think perhaps the *first* test should be with the code exactly as it is now, and then we can either prove that its something on our end, or it isn't...
Click to expand...

^^^
IIIIIII Good idea!
 
You may see some slowdowns today as I run some tests to try to figure out how long a full system copy will take.
 
Chuck,

Let me say thanks.


And Good Luck, we're all counting on you. :goofy:

James Dean
 
called godaddy during this last freeze - they didn't see anything.

What I could see was that all 4 proc's were tied up with > 90% iowait reports. :/
 
Greebo said:
called godaddy during this last freeze - they didn't see anything.

What I could see was that all 4 proc's were tied up with > 90% iowait reports. :/
Click to expand...
I had timeouts from 1100 to 1107 EDT.

Then again while attempting to post this. :p
 
Just another clue, but I see the same thing over at AOPA.

Might be something VB does periodically to the DB like an index optimize or garbage collection.

Joe
 
Last edited:
The freeze, or I should say A freeze, happened this evening, 10/16 at 6:40 PM and lasted about 5 minutes.
 
Areeda said:
Just another clue, but I see the same thing over at AOPA.

Might be something VB does periodically to the DB like an index optimize or garbage collection.

Joe
Click to expand...
to follow up on this I posted my experience on AOPA and evidently this is a common occurance over there also http://forums.aopa.org/showthread.php?t=35258

I did a quick search over at VBulletin.com/forum but didn't see anything there to explain it.

Joe
 
You must log in or register to reply here.
Back
Top