Double-Posts - Server Hang Issue
Double-Posts - Server Hang Issue
Ladies and Gentlemen:
Hmm, after saying that, I'm not sure we have many ladies...
None the less...
As many of you have noticed, f150online.com has been experiencing some server hang issues due to the load being placed upon our network due to increased membership and web-trotting bugs. With this in mind, I ask that you only click 'POST REPLY' or 'NEW THREAD' once instead of clicking it multiple times.
If you feel that your post did not go through, please back out, reload the page and confirm that it didn't go through before clicking 'POST REPLY' again.
This will help in two predominant ways.
1) It'll help keep threads clean, easy to read, and... most of all...
2) It'll make my life easier so that I'm not having to hop through and clean up the double, triple, and, yes, quadruple posts that I've been seeing here.
Thanks for your help folks. I really appreciate it!
RP
Hmm, after saying that, I'm not sure we have many ladies...

None the less...
As many of you have noticed, f150online.com has been experiencing some server hang issues due to the load being placed upon our network due to increased membership and web-trotting bugs. With this in mind, I ask that you only click 'POST REPLY' or 'NEW THREAD' once instead of clicking it multiple times.
If you feel that your post did not go through, please back out, reload the page and confirm that it didn't go through before clicking 'POST REPLY' again.
This will help in two predominant ways.
1) It'll help keep threads clean, easy to read, and... most of all...
2) It'll make my life easier so that I'm not having to hop through and clean up the double, triple, and, yes, quadruple posts that I've been seeing here.
Thanks for your help folks. I really appreciate it!
RP
I've waited as long as 5 minutes waiting for my post to go through. Sometime resubmitting, it goes right thru no problem and I don't get doubles often. If it didn't go in a minute and I re-hit reply, I'm usually good.
Dzervit,
That is the funniest damn thing I've read you say so far.
That is the funniest damn thing I've read you say so far.
Trending Topics
I'm curious why you think the hanging issue is directly attributable to server loads? I suspect bad code somewhere here and sometimes our browsers don't automatically dismiss it and get hung up trying to process it. I have to keep two separate browsers running and when one hangs, I access the site on the other browser and that seems to kick the hung window into gear. This isn't a server load issue, this is simply mis-written code within the website.
JMHO.
JMHO.
This is my guess as to what is causing the problems.
You guys use InFlow to host your site in a datcenter in Atlanta, right? I would bet that the datacenter is also hosting some Voice over IP applications.....and that the datacenter is using QoS (Quality of Service). QoS is a IP network standard that, in essence, gives priorty to certain kinds of network traffic. Voice packets will have priorty over, lets say, web server packets. So if on a given segment of the network, F150online's web server is cranking out the forum traffic, and along comes some voice data on the same segment, QoS policy says:
"Hey, sorry web traffic. The important data packets need to use the wires. You have to wait in this line to use the wires (when the voice packets are through with them)"
The idea is that voice traffic needs to be un-interupted, or else the quality of the calls is hindered. (voice can't wait because the call would be garbled)
However, a web server serving up web pages, it doesn't hurt to wait and re-send packets. But, in this case, maybe the web re-transmitions are not being handled well by the web server application. Or maybe the datacenter is selling more bandwidth to their customers than they really have (unheard of !!).
Just some thoughts, and I repeat, this is a guess !!! That said , check out this trace route...
You guys use InFlow to host your site in a datcenter in Atlanta, right? I would bet that the datacenter is also hosting some Voice over IP applications.....and that the datacenter is using QoS (Quality of Service). QoS is a IP network standard that, in essence, gives priorty to certain kinds of network traffic. Voice packets will have priorty over, lets say, web server packets. So if on a given segment of the network, F150online's web server is cranking out the forum traffic, and along comes some voice data on the same segment, QoS policy says:
"Hey, sorry web traffic. The important data packets need to use the wires. You have to wait in this line to use the wires (when the voice packets are through with them)"
The idea is that voice traffic needs to be un-interupted, or else the quality of the calls is hindered. (voice can't wait because the call would be garbled)
However, a web server serving up web pages, it doesn't hurt to wait and re-send packets. But, in this case, maybe the web re-transmitions are not being handled well by the web server application. Or maybe the datacenter is selling more bandwidth to their customers than they really have (unheard of !!).
Just some thoughts, and I repeat, this is a guess !!! That said , check out this trace route...
I'm fairly certain it is a MySQL issue.
What is likely causing the server to hangup is MySQL processes locking up while they are in queue because of either:
1) The search index has become too large (something vBulletin is known for because of its horrible use of a table to search words) and every write to create or edit a thread has to access that table
2) We've hit the point (once again) of needing additional hardware (likely memory) to handle the increased traffic load.
I am working on the first item now (and it will take a while) and the second item next week when I free up some available credit on the corporate AMEX.
All I ask is to please be patient.
added...It's amazing, all other things being equal, server load went from 0.80 to 0.27 when search was disabled...
What is likely causing the server to hangup is MySQL processes locking up while they are in queue because of either:
1) The search index has become too large (something vBulletin is known for because of its horrible use of a table to search words) and every write to create or edit a thread has to access that table
2) We've hit the point (once again) of needing additional hardware (likely memory) to handle the increased traffic load.
I am working on the first item now (and it will take a while) and the second item next week when I free up some available credit on the corporate AMEX.
All I ask is to please be patient.
added...It's amazing, all other things being equal, server load went from 0.80 to 0.27 when search was disabled...


