I had the chance to work with the summit24 switches, and I personally liked them. It is straight-forward in my opinion for the Web Interface compared to others I have seen and the pricing seems to be reasonable.
I was wondering if anyone has had any experience with any of their switching devices.
This is a instructional overview thread for those developers who are getting into setting up their own server with a LAMP (Linux Apache MySQL PHP) setup. The linux distro referred to in this thread is a centOS (fedora|redhat) setup.
- Make sure that your actual RAM is the same amount that is displayed by the server (there could be some BIOS restrictions on RAM so check for that).
Linux OS Installation
- Use a server system for the type of install. ------------------------------ - Setup you partitions with care: - Make sure that everything other than /swap is an ext3 partition type. - /swap (usually double the amount of space that your ram has but never larger than 4 gigs. - /tmp (700meg is ok). - /boot (100meg is ok). - / (leave rest of the available space in the harddrive for this). ------------------------------ - Use GRUB boot loader - Use DCHP only if your IP address changes due to the network. If not, then assign the IP address of the box. - Assign the netmask if DCHP is not in use. ------------------------------ - When setting up the packages, select only what you need. Most of the time it's better to just install no packages and then install everything you need by yum (yellowdog update manager). If you do not select any packages, only the 1st CD of the linux install will be needed.
Linux OS Customizing
- Create a new user and provide it a password (with # passwd). Do not create a user with a generic or commonly known names used in any daemon programs (ex. mysql, apache, admin, user, php, postgresql). - Disable the ROOT login in SSH (this means that when you login using the other user with SSH, you'll have to $ su to the root user). - Install "Development Tools" with yum using group install if you plan to compile your own apache. If not then install apache with yum install apache.
- Disable the extensions that you're not using for your website. If the server is only hosting one website, then there is no use for Virtual Hosts. - Set the ServerLimit value to a suitable value so that users won't get locked out of the website. - Change the User and Group directives to the newly created user. - Set Options +Indexes to Options -Indexes so that the contents of directories w/o an index file will not be displayed. - Change the DocumentRoot setting to the newly created user's home (~) directory. Or if you plan to use the default (/var/www/htdocs or /var/www/html) then assign the permissions of the user to that directory. - Add apache as a start up program when the operating system boots up (this can be set in /etc/rc.d/local). - Setup logs accordingly. If you setup image logging and your server has 20+ images per page then your website performance can suffer. - Setup Error Logs to a suitable level. - If any web pages are not displaying and the web server appears to be on when accessing it from the localhost (wget http://localhost) then disable or flush the iptables (/usr/sbin/iptables --fliush). You should also set this as a start up option for the OS. - Use mod_rewrite to use modern URLs.
- Use --skip-name-resolve. - Use --skip-bdb (if you're not using it). - Use --skip-innodb (if you're not using it). - Set a log for slow queries. - Set the max_connections to a high value. - Do not set a user with a wildcard ip-address. Only setup users with a specific ip. - Use Query Caching for frequently used queries.
- Disable Magic Quotes. - Disable Register Globals. - Disable Short Tags. - Disable ERROR REPORTING if the website is not in development mode. - Enable HTTP Only Session Cookies. - Set Session Cookies to only be cookies (and not URL's). - If sessions do not work, then set the session save path to a directory where the apache user has access to. - Use Gzip Compression.
- Use an optcode cache for PHP (Eaccelerator). - Consider using a static domain for CSS and JS files (this way the same cookies for the website won't be sent on each request). - If your website uses alot of CSS and JS files per page, bundle all of them together into one request using mod_rewrite and php [url] - For Apache, use the lingerD module (this reduces the amount of resources that are used when an apache connection is closed).
Here are some links for optimizing your server build:
Feb 19 15:57:39 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:06:02 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:06:02 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:14:24 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:14:24 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:22:46 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:22:46 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:31:09 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:31:09 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:39:31 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:39:31 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:47:53 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:47:53 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 16:56:16 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 16:56:16 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 17:04:38 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 17:04:38 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 17:13:00 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 17:13:00 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed. Feb 19 17:21:23 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session opened. Feb 19 17:21:23 server proftpd: server.com (127.0.0.1[127.0.0.1]) - FTP session closed.
I see over a few hundred of these lines in /var/log/messages. The timestamp is exactly the same for every 2 lines (Proftp session- Opened and Closed). It's occuring every hour of the day. Is someone attacking the ftp daemon or something?
Having a slight problem working with one of our Extreme Summit 48 (ugh) switches - I've figured out most of the basics, but I can't seem to find any way to add a secondary IP address to a VLAN! This, I would have thought, would be a pretty basic feature to have. Typing "config vlan [vlanname] ipaddress 220.127.116.11/24" works for setting the primary IP, but I can't figure out how to add any more - and doing the command again just overwrites the first one.
So... does anyone have any tricks up their sleeve, or is this something that Extreme neglected to add to this model switch?
My host tells me that they have security to stop DDoS attacks and stuff, however today my server load jumped to 17.12 and my site went down giving me a "Network Timeout" error.
My host tells me it's my fault that I am using too many resources. The MOST my site has been on load is 3.06 and that was around lunch time a few weeks back. It's 11:43 and the server load is 17.12? I think my host is pulling my leg. I have not added ANYTHING new to my site and have not changed anything in 3 days. The load has been fine till today.
I work for a company with a highly trafficked site (about 100,000 unique daily visitors). We have our "MANAGED" hosting through Rackspace for quite the pretty penny per month. It includes a dedicated back end / DB server and cloud servers for the front end.
The setup has been functioning perfect for 10 months now; but this past Monday the speed of the site immediately dropped. Page load times fell from 1-2 seconds to between 10-20 seconds, and sometimes not at all. As far as we know (and as far as Rackspace says), no server setting were modified. No new code was introduced on our end. It's a mainly static site, with minimal user interaction with the backend at all.
We've monitored the traffic, checked IPs, etc. We've even tunes down several site features in the interest of reducing server load. Upon a server reboot, the active threads/processes running on it IMMEDIATELY jump back up to maxed out levels. It seems like once our daily traffic reaches 10MB/s, a type of queue forms and the delays begin. Rackspace assures us that we're not limited to that.
Versions: OS: cent OS on cloud OS: Redhat on Dedicated Server Apache: 2.2 PHP: 5.3 / MySQL: 5.1.69
Some more background info: The site is typically busiest from 7am until 3pm EST. For the past few days, we've noticed that between 7am and 9-10pm the server has just lagged incredibly. However, at around that 9-10pm mark, something changes and the pages go back to loading almost instantly. (There is still decent traffic though.) Then at around 7am again it slows to a crawl.
Rackspace has offered solutions such as spinning up another server and incorporating their load balancing - they are in the process of this BUT they do NOT think the traffic is the issue. At one point they actually said there was potential packet loss somewhere in the network, but no progress has been made.
I'm having a hard time finding a source that has the Supermicro CBL-0084L splitter cable in stock. Any suggestions where I could get my hands on about 10 to 20 of these at a reasonable price? Maybe more if the price is right.
I have had a few Colo and Dedicated servers but I have decided that the traffic I am getting on my sites warrants using a large setup so I have four servers that I need to colo to one site.
My major question is that when I go colo they generally only supply one Cat 5 cable and what is the best setup to have for a switch/router? Generally you get enough IP addresses to have one per server so my guess is that you can just plug the switch/router in and away you go?
I am going to have one server as a load balancer then two as the load balanced pair and then a database server.
I've noticed on some rackmount servers and firewall appliances there are ports labelled console, and look like serial ports. Are these related to connecting to KVM? I am looking for a cheap appliance for some rackmount servers I have (1U half depth would be ideal), to remotely recover the server if it goes down. I figured they'd connect through PS2 connectors, but then I thought that if the console port was related, then going through that might have the option for power cycling or something.
I'm trying to knock a bug out of our new Hardy Heron Xen template. Basically when the old ones boot up /bin/sh is spawned on the console. Now with Hardy init has been replace with /etc/event.d handling.
So I've replace /etc/event.d/ttyS0 and ttyS1 (not sure which one is the console but I think ttyS0)
Code: # tty0 - shell
# This service maintains a getty on tty1 from the point the system is # started until it is shut down again.
start on stopped rc2 start on stopped rc3 start on stopped rc4 start on stopped rc5
stop on runlevel 0 stop on runlevel 1 stop on runlevel 6
I know typically at least CAT5 cable is to be used for FE or Gig-E Ethernet links, but can CAT3 cable be used at short distances? I know there is little to none cost difference in the cable price, but I am only allowed to used the preinstalled cable at this location which is CAT3 which is what Verizon typically runs for telephone cables.
So we have 10 racks of mess in our custom colo cage. Each rack currently has 20+ servers, most with dual network interfaces hooked up, most racks with two switches and 3+ apc remote power strips per rack. Each rack is an APC 4 post unit [url]
We have no cable management trays and currently tie up cables with velco and zapstraps, trying to keep power cables to the left and cat5 cables to the right. Unfortunately, even with all our effort, it's really turning into a wall of cabling that is basically insulating each rack and causing a temperature issue (or the potential for one at least).
Even worse, we want to add in KVM over IP hooked up on nearly every server. While I see the great benefits of the KVM, I can't see how on earth we are going to make the cabling work out!
I am upgrading my current colo'd server from a 1U dual core xeon with 2x500GB drives to a 2U core 2 quad with 8x750GB drives (raid5).So far I have ran into two issues. One is the rails are too large for the cabinets (problem #1 which I put in another thread). The only other problem is I am having problems with remote console on the new hardware.
It should be setup correctly since I basically rsync'd everything from the old server to the new one. The device is detected:
root@houkouonchi: 03:19 AM :~# dmesg | grep -i tty Command line: root=/dev/sda2 gpt pci=nomsi console=tty0 console=ttyS0,9600 notsc Kernel command line: root=/dev/sda2 gpt pci=nomsi console=tty0 console=ttyS0,9600 notsc console [tty0] enabled console [ttyS0] enabled serial8250: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A 00:05: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
But when I connect to the digi I am not getting any console output. I have tried a different serial -> Ethernet adapter and tried a different patch cable/port on the digi so I am pretty sure those are fine. The motherboard (nforce 650i chipset) didn't have an onboard serial port so I am using a port hooked up to a bracket which has a ribbon cable which connects to the motherboards COM port. I also tested another serial connector which the case has built in and both give me problems.
I put this in the co-location section since I am co-locating a server that I host a few Websites on. The server is located in the Pacfic Northwest at a hosting company out there. As you may know, this is a "more remote" area of the United States.
Today, the company lost all Internet access when a Fiber optic cable went down. Not only was my stufff down but the entire hosting company was down. In 10 years of doing Internet development, I have never seen this happen to any hosting company I have worked with no matter how good or bad they were. On top of that, it happened one time last year as well.
According to my co-location provider, the problem happened a long way up stream. SO far that the lines cannot be backed up. Is this true? Could a fiber optic cable fail at some point where it cannot be switched over to another line?
I don't know if I should I believe that or not. It would seem to me that it is a matter of money and they may not have a back-up system in place if the pipe goes down. Is it possible? Who's fault is it?