I upgraded my RaQ from 128k to 512k a few weeks ago because of high bursts of traffic every Sunday morning when I post exclusives on my sites. The upgrade absolutely did NOT solve my problem and all of the sites on the RaQ were virtually unaccessable for about 8 hours because of traffic.

Could anyone please tell me what I need to do to solve this problem! I'm looking like a pretty crappy host... ;)

More detail is needed on this one...

- are the accesses just of static content or is it database calls?

- how big are the files that are being retrieved? taking up alot of apache processes?

etc etc...

What company do you have your raq with...

If the company that you're getting the RaQ from has you on a capped line, it might be that you just don't have a big enough pipe to handle the data transfer. You can also specifiy (in the control panel) caps for different sites that your RaQ uses (based on IP#). Check to see if this was done accidentally.

<<<<- are the accesses just of static content or is it database calls?

- how big are the files that are being retrieved? taking up alot of apache processes? >>>>

First off, the RaQ is with Rackshack. RaQ4i, 20gigs space, 150gigs transfer/month, 512k.

The items being accessed are photos, about 6Mb. worth total. 100 or so at about 60k each. Absolutely nothing fancy at all http://shippers.fredfarm.com

Does this help any, or do you need more info? And thank you so much for helping me. :)

Originally posted by IntelligentHosting.com
If the company that you're getting the RaQ from has you on a capped line, it might be that you just don't have a big enough pipe to handle the data transfer. You can also specifiy (in the control panel) caps for different sites that your RaQ uses (based on IP#). Check to see if this was done accidentally.

OK, my RaQ is through Rackshack. I should ask if I am on a capped line? And as far as capping by IP address, I only have one IP address for my RaQ with about 10 sub-domains and 10 virtual sites. Or is there a way to cap by domain name?

You can cap by domain name as long as each domain name has a unique IP# (if you cap an IP#, you're capping all the virtual sites that share the same IP#).

Log on you your admin control panel, then click Control Panel, then Bandwidth. See if there are any bandwidth limits there.

You might also contact Rackshack and see if they're capping your line (although there's enough Rackshack people here that someone might know the answer).

Originally posted by IntelligentHosting.com
You can cap by domain name as long as each domain name has a unique IP# (if you cap an IP#, you're capping all the virtual sites that share the same IP#).

Log on you your admin control panel, then click Control Panel, then Bandwidth. See if there are any bandwidth limits there.

You might also contact Rackshack and see if they're capping your line (although there's enough Rackshack people here that someone might know the answer).

I checked and there are no bandwidth limits there.

I seem to remember when I was telneting that there was a section in there that asked something like "How many people can access the RaQ at one time" and the number was "300." Does this sound familiar to anyone? Because if it does, I need that number times about 1000! :)

You probably need to do some monitoring to see where the bottleneck is... It could be disk i/o, CPU, or software configuration. Without this information, it is very hard to troubleshoot your problem. But here are some things to check ....

Processes
Check your process list and kill any thing that is not necessary. Also, if any of your pages are called through scripts make sure that the scripts are dying off and not hanging around. We had this problem with a PERL script once. If you're not using chilisoft on your machine, kill it off. You will get back a few CPU cycles and some RAM. On production servers, we shut down everything that we do not use.

httpd.conf
Tweak httpd.conf to give you more spare servers and max clients. I have seen slow access times caused when there are not enough spare servers to handle burst traffic or the number of max clients is set too low. If Max Clients is too low, then people will have to wait and may even get server not found errors. If you have enough RAM, try setting your min spare servers to 20 or so and the max clients to 100. After making these changes, you will want to monitor your RAM usage. If you still have plenty of RAM, then add another 50 MaxClients.


Disk I/O
Unless you are pumping our a lot of data during burst, I doubt this is your problem. Most drives in Raqs, when properly configured, give you 20MB/sec or more read/write access speeds. And if people are accessing the same content, it should be pulled out of the cache.


My $0.02
I would think that you would have to have a lot of traffic to bring a Raq4 to its knees if you are serving up static content. We ran a web server with 50 virtual sites, some of them with moderate traffic, on a PII300 for years without a problem. However, the box was highly tweaked for delivering specific types of content and had a large number of MaxSpareServers and Clients to handle bursts.

http://thingy.kcilink.com/modperlguide/performance/index.html

This might help...check out the http://thingy.kcilink.com/modperlguide/performance/Performance_Tuning_by_Tweaking_A.html
Section.....

Thank you so much for the information and link huck! I am on my way to read it.