Archive for the 'General' Category

Existing SingleHop Clients- Here’s $100,000+ :)Zakby24 Jul 2008

admin

Hardware is constantly outdated which, normally, presents a big problem for dedicated hosting clients. You get a server today and in 6 months your provider will naturally be offering a replacement of your server for new clients- with bigger and better stuff.

While we can’t entirely solving that problem, we decided recently to begin only offering a minimum of 2GB of RAM on new servers. To show our commitment to our very loyal clients, we are providing a free RAM upgrade to bring each and every client to 2GB of RAM.

As far as we know, this is the first time a hosting provider has gone back and mass upgraded it’s clientèle on this level. Here’s the bottom-line impact and the reason providers don’t normally do this:

1,200+ servers potentially impacted (We feel in the area of 50% of these effected).
We’re also going to be reducing approximately 20% of our customers bills that pay for 2GB of RAM.
In all, over the course of a year, hardware expense and loss of revenue: Well over $100,000.

Sounds meek, right? To an accountant it might not be the best move, but our business is experiencing exponential growth. That’s largely based on our very loyal clients and reduced advertising costs because so many of you are talking positively about our services on various forums and websites. We’re showing our appreciation for your business and we would love to hear your feedback to this giveaway, or feedback on our services!

With the announcement of LEAP, we’ve had some extremely exciting times at the company. That’s going to get better with more announcements next week as we continue to pass up the competition in terms of new features and services all aimed at giving you more control over your infrastructure and done at affordable prices. (or in this case, an upgrade for free!)

Overall, we’re happy to be in the position to do it, and we hope that you appreciate the efforts and service that you receive from SingleHop.

To take advantage as an existing client, go here: http://www.singlehop.com/servers/free_ram_upgrade_offer.php

Read the full post "Existing SingleHop Clients- Here’s $100,000+ :)" 6 Comments

announcing LEAP — the server manager webtop (only from SingleHop)Dan Ushmanby21 Jul 2008

Dan

Hello Everyone,

This is a big post and I’m very excited to be making it. Over the past year we have been developing a new control panel – a new customer portal. The system is finally ready for release and I wanted to explain why it is cool and some of its major advantages.

LEAP is entirely browser based. It is tested on Windows, Macintosh and Linux using Opera, Mozilla, Internet Explorer and Safari. The system is a graphical, Windows-like interface for managing your servers. But, it’s more than just a pretty interface, it’s a fully functional control panel – allowing you to add servers, managing existing servers, and administer those servers remotely using built in SSH and RDP clients. The system features automatic server provisioning technology as well as advanced graphing and management features. In a nutshell, it’s what every dedicated server customer dreams about – a dream that has gone unanswered by just about everyone. It’s a dream that we’re answering now with the first public release of LEAP.

To watch a cool video (less than 5 minutes long) and learn everything you can about LEAP, click here:
http://www.singlehop.com/poindexters_lab/leap.php

If you have any questions or comments, please post them in this thread.
Dan Ushman

Read the full post "announcing LEAP — the server manager webtop (only from SingleHop)" 1 Comment

Hardware vs. Software FirewallsGarett Murphyby16 Jul 2008

By default most computers are left vulnerable to attack the moment they are attached to a network because they either have an inadequate firewall or no firewall configured on their machine. This is the case, in many situations, for both servers and your own desktop at home. It is for this reason that I am often asked the question “How good is an operating system-based firewall”. The answer is different for each situation.

In the home situation, it can be used as your primary firewall (for example, you do not have a router between your DSL, cable, or FiOS modem hardware) provided you have configured it correctly (or at least allow the operating system to manage it for you). In a dedicated server hosting environment, however, it is extremely important to have a firewall of some type enabled and to manually configure it to filter your incoming traffic.

Read the full post "Hardware vs. Software Firewalls" 1 Comment

Backups Backups Backups!ericby07 Jul 2008

It happens all too often. I client calls us and tells us they just deleted a folder or a file that was extremely important, usually for a presentation that is to be made in no more than 30 minutes. Now, if they have backups on our backup servers, that is fine, we restore the file in no time because they are available right there. This isn’t the normal situation though, most people don’t feel that making backups is important, or think that we just do it because they have the server with us (we don’t). If these backups don’t exist, and your server crashes, gets hacked, explodes, runs away, or heaven forbid a natural disater occurs, your data is gone. There is no way around it. That is why making sure you have backups of all of your data is so important.

The first thing thing you need to do is figure out where you are going to store backups. You can store them in one place, and that would be fine, but it also leaves a single point of failure. I would recommend keeping your backups in at least two places, perhaps keeping a copy on your home computer, and one at a remote backup server. You could also just use two remote servers in different physical locations. We do have backup servers in our datacenter, just ask your account executive if you are interested.
After you have figured out where you want the backups to go, you should figure out what data is important enough to you to backup. The best way to go about this is to just backup everything short of the operating system itself. Get all of your web content, settings, databases, mailboxes, dns zone files, user accounts, and the list goes on.

With cPanel, this is easy. There is a built in backup function that will automatically backup all of your stuff daily. If something were to happen, all you would need to do is move the backup files to a new cPanel server, and restore them. It will restore everything back to it’s previous functionality relatively quickly. Windows, with or without a control panel, is pretty easy as well. It has a program called ntbackup that you can schedule backups with, all you need to do is choose the location, the files and folders to backup, and when to do it. It will do the rest for you. If you run a Linux server without a control panel, setting up backups can be a bit more complex without a backup program. You can use a mixture of tar, cron, and scp or ftp to run backups when you would like. If you need help with something like this, just let us know, we would be glad to help.

I cannot stress this enough, making backups of your data is absolutely vital. If you use your server to host sites or applications that bring you money, not having backups is like asking to lose it all in a second. If you keep backups, you may have some downtime, but at least what is important to you can be recovered.

Read the full post "Backups Backups Backups!" 1 Comment

Spanning Tree Protocol Explained!Aziby04 Jul 2008

Spanning-Tree Protocol (STP) is a link management protocol that provides path redundancy while preventing undesirable loops in the network. When it comes to Ethernet networks, in order for them to function properly only one active path can exist between two stations. Loops occur in networks for a variety of reasons. The most common reason you find loops in networks is the result of trying to simply provide multiple levels of redundancy. When having multiple root switches in a network used to provide this redundancy, the problem that will arise is infinite loops between the switches and the end workstations they’re connected to. If a loop exists in the network topology, the potential exists for duplication of messages. When loops occur, some switches see stations appear on both sides of the switch. This confuses the forwarding algorithm and allows duplicate frames to be forwarded.

Spanning-Tree comes in to play by only allowing a singular path to destinations when having multiple (redundant) links such as this. In a STP environment, the switches exchange information amongst themselves using bridge protocol data units (BPDU) and will then listen in on all ports for this BPDU message. Once a bridge is turned on, it automatically assumes that it is the root bridge in the STP tree. The STP software chooses a root bridge and calculates all paths from the lower bridges back to itself. In the event of hardware failure of a root bridge in the redundant environment, a new root is elected and port paths would be recalculated.

Here at SingleHop we currently utilize three and very soon to be 4 uplink providers. With this uplinks STP is a must in our switched environment to provide multiple redundancies in case one goes down, another link is there to takes it place within a matter of seconds. At this stage in our continuously growing network topology there’s virtually no way you’ll ever lose the connection to your dedicated server. That is of course you don’t reboot the server itself, then that’ll be your expected momentary loss of connectivity.

Read the full post "Spanning Tree Protocol Explained!" No Comments

MySQL Tuning-Primer.sh!Andyby03 Jul 2008

Andy

The most important aspect of any MySQL-Driven web application is tuning MySQL to perform at it’s peak. Typically the SingleHop Technicians do this all manually, which takes a lot of time and perseverance. However, recently Igor from the Development Team introduced this very easy to use script — mysql-tuning-primer.sh

The script was developed by Matthew Montgomery from MySQL Inc. — not bad for a “Novice” MySQL Technical Support Specialist. I’d love to see what MySQL’s Senior technicians can do! Anyway, onto the script…

Currently the script handles recommendations for the following:

  • Slow Query Log
  • Max Connections
  • Worker Threads
  • Memory Usage
  • Key Buffer
  • Query Cache
  • Sort Buffer
  • Joins
  • Temp Tables
  • Table (Open & Definition) Cache
  • Table Scans (read_buffer)
  • Table Locking
  • Innodb Status

The script can be easily downloaded to any linux-based server running MySQL, and is Plesk and WebMin friendly!

function check_for_plesk_passwords () {

## — Check for the existance of plesk and login using it’s credentials — ##

if [ -f /etc/psa/.psa.shadow ] ; then
mysql=”mysql -S $socket -u admin -p$(cat /etc/psa/.psa.shadow)”
mysqladmin=”mysqladmin -S $socket -u admin -p$(cat /etc/psa/.psa.shadow)”
else
mysql=”mysql -S $socket”
mysqladmin=”mysqladmin -S $socket”
fi
}

function find_webmin_passwords () {

## — populate the .my.cnf file using values harvested from Webmin — ##

cecho “Testing Stored for passwords:\c”
if [ -f /etc/webmin/mysql/config ] ; then
user=$(grep ^login= /etc/webmin/mysql/config | cut -d “=” -f 2)
pass=$(grep ^pass= /etc/webmin/mysql/config | cut -d “=” -f 2)
if [ $user ] && [ $pass ] && [ ! -f ~/.my.cnf ] ; then
cecho “Setting login info as User: $user Password: $pass”
touch ~/.my.cnf
chmod 600 ~/.my.cnf
echo -e “[client]\nuser=$user\npassword=$pass” > ~/.my.cnf
cecho “Retrying login”
is_up=$($mysqladmin ping 2>&1)
if [ "$is_up" = "mysqld is alive" ] ; then
echo UP > /dev/null
else
second_login_failed
fi
echo
else
second_login_failed
echo
fi
else
cecho ” None Found\n” $boldred
second_login_failed
fi
}

After downloaded, simply run it from the command line and read the output:

Looks like I have some work to do! Until next time, happy hosting!

Read the full post "MySQL Tuning-Primer.sh!" No Comments

HOWTO: add drives to your Linux/cPanel serverAndyby01 Jul 2008

Andy
Hi I’m Josh the new guy here at SingleHop, and today I’m going to be explaining how to add additional hard drives to a Linux/cPanel server. Before I get into that, I’ll give everyone a little background information on myself. I’ve been working in the webhosting field for quite a long time now doing everything from answering phones, building servers, and taking out the trash. My very first job for a now defunct hosting company was back in 2000. We ran Sun Cobalt RaQ server appliances and a few cPanel v3 servers. After my days were up at my first employer, I continued to do systems administration work for various companies and individuals, both on site and remotely. In late 2006 I was hired by another local dedicated hosting company located in Arlington Heights, IL. My job was to build servers, and pretty much take care of them. During this time I was able to really fine tune my skills with Linux and cPanel. Fast forward to today, and here I am down at SingleHop doing the exact same thing, looking after servers. I usually try to spend my off time doing something relaxing, hanging out with friends, stuff that keeps me pretty far away from the computer.

Okay, so. Your reading this to figure out how to add another hard drive to your cPanel system running some version of Linux. For my purposes here we’ll assume we’re all using CentOS 5.2. First things first, before anything is added/removed from your server, it’s a pretty good idea to take a full backup of everything and save it locally! In the past few years of doing this I’ve learned that it pays to keep backups.

Once your backup process has been completed and your new hard drive is added we should be all ready to go right? Wrong. We’ll first need to format/partition the drive and then mount it.

Note: Hard drives are identified by interface. IDE hard drives are prefixed with “hd” and SCSI/serial drives are prefixed with “sd.” You can use letters a-h for IDE disks, and letters a-p for SCSI/serial drives. Some distros do have their own special letter schema’s, but are usually well documented if they are.

We’re going to assume we’re installing a new serial ATA hard drive into our system, and it’s hooked to the second SATA port, thus making our drive “sdb.” First things first log into your server as the root user. Once you’ve done that enter the following as root:

Code:
# fdisk /dev/sdb

This brings up how many cylinders are on your hard disk, and of course opens up the fdisk program that will allow us to partition our new hard disk.

The next step is to simply enter a “p” to show the partition table. This is useful to see what partitions are already setup on the disk, if any.. The output will look something like this:

Code:
Command (m for help): p

Disk /dev/sdb: 50.0 GB, 50019202560 bytes
255 heads, 63 sectors/track, 6081 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot    Start       End    Blocks   Id  System

Command (m for help):

Now, since my drive here is a brand new drive, there are no partitions on it. If any partitions are present we’ll just go ahead and delete them by pressing “d.”

Now everyone should have the same thing. An unpartitioned hard drive. Next we’ll need to partition our drive. We’ll keep it simple and mount this drive with a large, single primary partition, just like a backup drive. To do that enter “n” at the command line to create a new partition. You should see something like this:

Code:
Command (m for help): n
Command action
   e   extended
   p   primary partition (1-4)
p
Partition number (1-4): 1
First cylinder (1-6081, default 1): 1
Last cylinder or +size or +sizeM or +sizeK (1-6081, default 6081): 6081

Command (m for help):

As you can see, I hit “n” and it asked if I’d like to create a primary (P) or an extended partition (E) I hit P for a primary partition, entered “1″ for it being my first partition on the drive. Your first cylinder will be “1″ and the last one in our case will be “6081.” Most times you will be able to hit enter and use the default cylinder counts.

Let’s verify that we’ve done everything correctly up until this point. At the command prompt enter “p” to check the partition table again:

Code:
Command (m for help): p

Disk /dev/sdb: 50.0 GB, 50019202560 bytes
255 heads, 63 sectors/track, 6081 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot    Start       End    Blocks   Id  System
/dev/sdb1             1      6081  48845601   83  Linux

Command (m for help):

Now you can see that we’ve added a partition to the drive, but we’re not done yet. No changes have been written to the disk yet, everything is just in memory. If everything looks right, go ahead and enter the “w” command which will write the changes to the disk.

Now that our disk is partitioned to sdb1 we’ll need to format it. You can format your drive to any filesystem you want, but for our purposes here I’m going to format it with ext3, the most common today. To format your drive:

Code:
[root@chicago]# mk2fs -t ext3 /dev/sdb1
mke2fs 1.32
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
6111232 inodes, 12211400 blocks
610570 blocks (5.00%) reserved for the super user
First data block=0
373 block groups
32768 blocks per group, 32768 fragments per group
16384 inodes per group
Superblock backups stored on blocks:
        32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
        4096000, 7962624, 11239424

Writing inode tables: done
Creating journal (8192 blocks): done
Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 38 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.
[root@chicago]#

Hmmm? What all did we just do? Using the mk2fs (make file system) command, we specified the type (using the -t) ext3 using the device and partition name (/dev/sdb1). You have successfully partitioned and formatted your new drive, but we’re still not done. We need to mount this partition to make it usable, and add an entry in the file “/etc/fstab”

We’re going to add this entry in the fstab file because on reboot, our drive that we just mounted will need to be remounted. By adding a simple entry in the fstab file our drive will be mounted every time on startup. To do this we’re going to add the following code on the last line of the file “/etc/fstab”

Code:
/dev/sdb1               /backup                  ext3    defaults        1 2

Once you’ve added that to the last line in the file, save it and issue this final command as root:

Code:
# mount /dev/sdb1

The mount command mounts the drive for immediate use, and the /dev/sdb1 is our drives partition name.

That’s it! You’ve just added a new hard drive to your cPanel server the “hard” way. cPanel comes with a built in handy utility that will automatically detect, format, and partition your new drive with just a few clicks.

To install your new hard drive through cPanel you’ll need to login to WHM as root, and find the button on the left called “Format/Mount a new hard drive” under the drives tab about 3/4 the way down. Click on that. The next page will show you the newly detected drives and have a button next to the drive to select it. Once you select your new drive follow the onscreen instructions, and hit enter. In just a few minutes WHM takes care of everything for you.

Edit/Delete Message

Read the full post "HOWTO: add drives to your Linux/cPanel server" No Comments

Installing ASP.NET 1.1 with IIS7 on Windows 2008Samby30 Jun 2008

It seems that with each new release of Windows system administrators have the world’s worst time migrating from the previous operating system release to the latest. One of many compatibility issues with 2008 is that .NET is built into Windows 2008 but is only 2.0, 3.0 and 3.5. When you try to install ASP.NET Windows warns that there are compatibility issues and no solution to installing this software on your operating system. So after doing some searching I found the installation steps and tried it out. This is the method I used to install ASP.NET 1.1 on Windows 2008 Web:

First you have to install IIS Metabase Compatibility. This is done by opening the Server Manager from your start menu. Once the Server Manager has loaded look on the left-hand side, click on Manage Roles and then Web Server (IIS). Check the box next to IIS Metabase Compatibility which is located under Management tools, IIS 6 Management Compatibility.

Now it is time to install ASP.NET 1.1. First download these three packages and run them. You will receive a warning that it won’t work. The reason this message comes up is because these applications are 32-bit only and IIS7 runs in both 32-bit and 64-bit. So if you tried to mix the 64-bit application pools with 32-bit application pools your application would crash, hence the compatibility issue. You have to install all 3 of these ASP.NET packages or it will crash on your system. This is because Windows 2008 has Data Execution Protection and the older ASP.NET is not compliant without Service Pack 1.

Now you must enable the ASP.NET ISAPI extension. To do this open a command prompt as administrator and run “%windir%\Microsoft.NET\Framework\v1.1.4322\aspnet_regiis –enable”. Easy enough right?

Now you have to edit the machine.config or else ASP.NET will throw runtime exceptions. To do this open %windir%\Microsoft.NET\Framework\v1.1.4322\config\machine.config in notepad as administrator and add the following code just above </configSections>:

<section name=”system.webServer” type=”System.Configuration.IgnoreSectionHandler, System, Version=1.0.5000.0, Culture=neutral, PublicKeyToken=b77a5c561934e089″ />

Save the file and then move your domain to the ASP.NET 1.1 application pool via the IIS manager. Once you have done this your website will be using 32-bit ASP.NET 1.1. I hope this has been helpful as this turned out to be quite an annoyance for me. While transitioning to Windows 2008 seems to be quite difficult I feel that things will get easier as time progresses. I just wish Microsoft had put more care in supporting their older technologies to ease the transition.

Read the full post "Installing ASP.NET 1.1 with IIS7 on Windows 2008" No Comments

FireFox 3: What with all the hype?luisby26 Jun 2008

Firefox is our favorite browser here at Singlehop, as it is for millions of other users and now its gotten even better! On June 17th Mozilla officially launched firefox 3 and 24 hours later it was downloaded over 21 million times from all around the globe. If you aren’t one of those who have downloaded the newest version and enjoying the improvements, let me fill you in on some of the newest features.

1. ITS FAST! - In previous versions Firefox’s memory used to grow out of control when you added tabs, extensions etc, but not anymore! The dev team’s primary goal was memory reduction. You can read the details on one of the developers’ blogs, Firefox 3 Memory Usage.
From just going to the sites I normally visit I can feel how blazing fast it is but for more evidence I used the sunspider javascript benchmark tool. I ran it in Firefox 2 first and then started it on Firefox 3. Firefox 3 finished before 2 (IE took twice as long as firefox 2) and the test results showed why. But here check out yourself http://webkit.org/perf/sunspider-0.9/sunspider.html
2. Improved Security – Firefox 3 has new security features such as Instant Website ID and Anti-Malware Scanner. You can check the validity of any site you go to by clicking on the favicon ( the icon just to the left of the URL in the location bar). This shows whether or not you are on the site you expect to be on and that the site owners are who they say they are. The three possible colors are

Grey – the site provides no identity information, which is fine if theres no sensitive information passing through the site. So most pages will have a grey icon.
Blue - Basic identity information shows that your connection to the site is encrypted and the domain has been verified and is not being spoofed. But, theres no guarantee as to who owns the domain in question. If you are still concerned. Click on More Information to open the privacy panel, here you can see the site’s identity certificate, if you’ve visited the site before, and if you have any cookies or passwords stored for the site.
Green - Complete identity information Go to paypal.com and you will see part of the location bar turn green. When the Site Identity button is green, that indicates that the site provides fully verified identity information about its owner, and that the connection is encrypted. This site is using the new Extended Validation (EV) certificate, which is much more rigorous in verifying a site’s identity than other certificates. The Green Icon says this connection is encrypted and the site is what you expect it to be.

3. Anti-Malware - If you happen upon an malicious website. Firefox 3 will popup a full-sized message as a warning against viruses, worms, trojan horses and spyware.4 .One-Click bookmarking – The built in list of attack-sites is constantly being updated and shows when to stop you from browsing, so there’s nothing for you to update. Other security enhancements include; insecure versions of plug-ins won’t run, Use of the Microsoft Antivirus API for full integration with antivirus software, and it also works with Vista’s parental controls, so you can better control where your kids go on the net.
At the end of every URL you will see a star. By clicking on the star firefox will bookmark that site for you. Two clicks on the star will open the bookmark manager allowing you to organize and tag the bookmark.
4. Smart Location Bar (aka The Awesome Bar) - The firefox location bar used to be just the place where you typed in a websites URL, but not any more. Now the location bar learns as you use it and adapts to your preference and improves its findings, to give you a quick way to get to the sites you love—even the ones that are just on the tip of your tongue. Type in a term and the autocomplete function includes possible matching sites from your browsing history, as well as sites you’ve bookmarked and tagged in a drop down. For example, you could enter the tag: “dedicated servers” to find “www.singlehop.com”. Matched terms are highlighted, making the list of results easy to scan.

If you are upgrading from Firefox 2 you may find some of your favorite addons are yet compatible with Firefox 3, but Firefox periodically checks for updates to addons and with so many people installing this version, you wont have to wait too long for updates.

Another little function has returned to firefox is the ability to block third party cookies. In order to protect you from advertisers who are collecting personal information, surfing behavior, interests etc
for more information on this check out http://www.grc.com/cookies.html

If you want to disable 3rd party cookies in Firefox heres how to do it. Open up Preferences(OS X) or Options (Windows & Linux), click on Privacy and then uncheck Accept Third Party Cookies. click OK and you are done. No more third party cookies.

So Firefox 3 has some great improvements and protective features and is a definite download must but always remember that you are the first line of defense when it comes to your computers security and that no software is FULLY SECURE, because just 24 hours after the release of firefox 3, security researchers have already found vulnerabilities. For more information check out. http://www.securityfocus.com/brief/759

Read the full post "FireFox 3: What with all the hype?" 1 Comment

Performance Tuning Apache the SingleHop Wayluisby17 Jun 2008

Apache 2.x is designed to balance flexibility, performance and portability. Apache is a good all-purpose webserver. Since Apache is designed to fit most scenarios it hasn’t been optimized to set any kind of speed records, but Apache 2.x is capable of high performance.

There have been many improvements made in Apache 2.x and many are enabled by default. But, there are some changes you can make at compile-time and run-time that can positively affect performance.

Ram is the single biggest hardware issue that determines webserver performance, so the more the better. Once a server runs out of memory it starts using the swap space on the hard drive. You never ever want have to the web server use swap. Using swap slows down of each request to such slow speeds which causes users to hit “reload”, which increases the work swap has to do, slowing it further.
So I always recommend putting in as much ram as your server can hold.

The rest of the hardware is up to you: get a fast CPU, a fast network card, and fast hard drives, you’ll have to figure out what is fast enough for you based on needs, availability and price.

You can counter, and should, control the MaxClients setting so that your server does not spawn so many children it starts swapping. This procedure for doing this is simple: determine the size of your average Apache process, by looking at your process list via a tool such as top, and divide this into your total available memory, leaving some room for other processes.

So lets go through some of the simple httpd.conf changes you can make to tune apache for higher performance.

Timeout - Is the number of seconds before receives and sends time out. By decreasing it to 15 seconds we can offset effects of a denial of service attack/server load

KeepAlive - This allows multiple requests to be sent over the same connection, which is useful when serving HTML pages with a lot of images. So make sure KeepAlive is turned ON. eepAlive

MaxKeepAliveRequests – This is the maximum requests allowed during a persistent connection. Zero allows an unlimited amount, so use a high number, for maximum performance

KeepAliveTimeout – This determines how long the server waits for the next request, put this on at a low value between two to five seconds

StartServers - This controls the number of child-processes that Apache spawns before accepting connections. Put this StartServers and MinSpareServers on high numbers,so if there’s a high load after restarting the server, there will be plenty of servers ready to serve requests.

MaxSpareServers and MinSpareServers – This tells apache how many processes to keep waiting for requests. Set this number high so apache wont have have to create new child processes to serve requests.

MaxClients – This is maximum simultaneous requests that the server can support. This number should be tuned very well. Too low and new connections end up in a queue and eventually time out, leaving unused resources. Set it too hight and the server will start using swap space and the response time with drop drastically.
So the trick is to properly calculated.

MaxClients = Total RAM dedicated to the web server / Max child process size

However if you need to configure more than 256 clients, you will have to edit the HARD_SERVER_LIMIT entry in httpd.h and recompile Apache

Here is an example of the numbers you should be using.
Timeout 15
KeepAlive On
maxKeepAliveRequests 100
KeepAliveTimeout 3
StartServers 8
MinSpareServers 10
MaxSpareServers 20
MaxClients 250
MaxRequestsPerChild 0

There are other things you can do to speed up apache but we will go over that next time.
If you have any problems setting this up on your server please submit a ticket to your support team at control.singlehop.com. We are here to help.

Read the full post "Performance Tuning Apache the SingleHop Way" No Comments

« Previous PageNext Page »