Config Recommendations Request: 25 High Traffic Domains
Mar 25, 2009
We are considering a new dedicated server host for a set of 25 domains, about 5 of which are very high traffic (80 million clicks a day each).
I am told that the biggest speed boost and performance comes from memory and fast hard disk. So I'm looking for at least a 16GB RAM and SCSI 10k 300GB hard disks.
I am also told that PHP etc is okay, but MySQL is the one that hogs resources after a while. So for the database server I need a high end server.
I like WHM/Cpanel so that should be alright.
A lot of email alerts may be sent through our system (about 500,000 a day on a good day as there's user subscription to updates functionality) so we may need a separate mail server?
My question: is the following config of 4 x quadcore Dunnington Intels be good enough for the above site? Can I run a database on this config for my kind of traffic or do I need a separate server too?
I'd like to get your recommendations on how I should approach this problem. I posted this in another part of the forum, but I feel that this is a more appropriate place for it.
Problem:
How to handle large amounts of traffic with for a social network website? If a user uploads a photo or video, how does it become accessible on all of the server? If traffic is expected to be about 500,000 visitors a day, how many machines do you think I should use?
Possible Solution:
I've come up with the following possible infrastructure.
One load balancer. The load balancer has 3 PHP/Apache servers behind it. Behind each of the PHP/Apache servers is a (slave) MySQL server, from which data is read. Behind the slave MySQL servers, there is 1 master MySQL server, which handles all of the database writes. The master MySQL and slave MySQL servers are synced up, so data is up to date.
The actual photo and video files are not stored in the database, only the links to them is stored in the database (to keep the database small). The photo and video reside in a central location (like a SAN or NAS), which is accessible by all of the 3 PHP/Apache webservers.
Questions:
1. How many machines do you think will be able to handle photo and video uploads for 500,000 visitors a day?
2. Is having a SAN with Terabytes of RAIDED disk space an available option?
3. If a SAN or NAS is not an option, does anyone have any ideas on how to make sure all of the web servers have access to the same photos and videos? Is rsync a viable solution?
4. Which hosting provider do you think I should go with?
5. Is clustering what I need? What is clustering and how will it address my concerns?
I'd like to get your recommendations on how I should approach this problem.
Problem:
How to handle large amounts of traffic with for a social network website? If a user uploads a photo or video, how does it become accessible on all of the server? If traffic is expected to be about 500,000 visitors a day, how many machines do you think I should use?
Possible Solution:
I've come up with the following possible infrastructure.
One load balancer. The load balancer has 3 PHP/Apache servers behind it. Behind each of the PHP/Apache servers is a (slave) MySQL server, from which data is read. Behind the slave MySQL servers, there is 1 master MySQL server, which handles all of the database writes. The master MySQL and slave MySQL servers are synced up, so data is up to date.
The actual photo and video files are not stored in the database, only the links to them is stored in the database (to keep the database small). The photo and video reside in a central location (like a SAN or NAS), which is accessible by all of the 3 PHP/Apache webservers.
Questions:
1. How many machines do you think will be able to handle photo and video uploads for 500,000 visitors a day?
2. Is having a SAN with Terabytes of RAIDED disk space an available option?
3. If a SAN or NAS is not an option, does anyone have any ideas on how to make sure all of the web servers have access to the same photos and videos? Is rsync a viable solution?
4. Which hosting provider do you think I should go with?
5. Is clustering what I need? What is clustering and how will it address my concerns?
We just upgraded to VBulletin 3.6.5, and are experiencing strange behavior. My forums are on a dedicated server, 27,000 users.
We'll be running along fine with loads of .5-2.00, etc. Then suddenly, the loads start climbing to 50, 70, 85+. We've been trying to figure out why. We even went on a different (much more robust) server, still the same result. This only started happening after the upgrade. Restarting the server corrects the problem, but only until it happens again. It can happen at anytime - during peak or off peak. The server may run for an hour or two until this happens, or it can happen 10 minutes after a restart.
Late last night when loads were normal we benched the server hard, and it ran just fine. We just can't figure out where this load spike is coming from.
Server info below:
Servers: Mysql: server version: 5.0.27-log PHP: PHP 5.1.6 (cli) Zend Engine v2.1.0, Apache Server version: Apache/2.2.3 Kernel: 2.6.19-1.2911.fc6 #1 SMP
Server: Pentium III with 1 gig of ram 15k RPM SCSI Raid 5 1 Gig RAM
cpu speed/type single or dual cpus): Intel Xeon 2x3.6 how much memory installed: 2048MB hard drive type/configuration: 2x160GB SATA linux distributor or windows version: CentOS 4.5 x86_64 apache/IIS version: apache 1.3.37 PHP version: php 4.4.7 MySQL version: 4.1.22-standard
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 19017 teknoloj 16 0 0 0 0 Z 27 0.0 0:00.93 php <defunct> 17751 mysql 15 0 241m 55m 3148 S 25 2.8 0:53.73 mysqld 19025 root 18 0 16204 6104 1392 D 6 0.3 0:00.17 cpcpan_check_in 18968 sohbetgo 16 0 44596 7540 3556 S 5 0.4 0:00.44 php 18552 nobody 15 0 32636 4208 1204 S 2 0.2 0:00.28 httpd 18808 resimsit 17 0 46944 9960 3616 R 1 0.5 0:00.23 php 18896 resimsit 16 0 46944 9984 3616 D 1 0.5 0:00.08 php 18921 resimsit 17 0 46404 9444 3572 R 1 0.5 0:00.09 php 18819 nobody 17 0 32644 4176 1192 S 1 0.2 0:00.06 httpd 18868 haylazt 16 0 0 0 0 Z 1 0.0 0:00.48 php <defunct> 18871 nobody 15 0 32504 4056 1148 S 1 0.2 0:00.06 httpd 18892 root 16 0 6416 1180 756 R 1 0.1 0:00.24 top 18978 nobody 16 0 32504 4028 1132 S 1 0.2 0:00.02 httpd 11 root 5 -10 0 0 0 S 0 0.0 0:24.56 events/1 2638 root 0 -20 0 0 0 S 0 0.0 0:16.96 loop0 2981 named 18 0 88016 2328 1152 S 0 0.1 8:44.12 named 3592 mailman 16 0 50316 2320 1828 S 0 0.1 0:00.78 python2.4
MySQL 4.1.22-standard uptime 0 0:39:9 Sat Jun 30 21:33:29 2007
__ Key __________________________________________________ _______________ Buffer used 19.34M of 80.00M %Used: 24.17 Current 33.70M %Usage: 42.12 Write hit 32.57% Read hit 98.45%
__ Tables __________________________________________________ ____________ Open 486 of 1800 %Cache: 27.00 Opened 702 0.3/s
__ Connections __________________________________________________ _______ Max used 12 of 500 %Max: 2.40 Total 4.33k 1.8/s
__ Created Temp __________________________________________________ ______ Disk table 26 0.0/s Table 2.23k 1.0/s File 428 0.2/s
6-7 vbulletin sites-vbseo and other scripts online avarage; 120 registered user ve 500+ unregistered user cookie timeout : 7200 unique hit : 15000-20000
KeepAlive ON MaxKeepAliveRequests 100 KeepAliveTimeout 2 MinSpareServers 10 MaxSpareServers 15 StartServers 10 MaxClients 180 Maxrequestsperchild value 1000
[root@server ~]# uname -a
Linux server.xxxxx 2.6.9-55.ELsmp #1 SMP Wed May 2 14:04:42 EDT 2007 x86_64 x86_64 x86_64 GNU/Linux
[root@server ~]# ulimit -aH
core file size (blocks, -c) 1000000 data seg size (kbytes, -d) unlimited file size (blocks, -f) unlimited pending signals (-i) 1024 max locked memory (kbytes, -l) 32 max memory size (kbytes, -m) unlimited open files (-n) 4096 pipe size (512 bytes, -p) 8 POSIX message queues (bytes, -q) 819200 stack size (kbytes, -s) 8192 cpu time (seconds, -t) unlimited max user processes (-u) 14335 virtual memory (kbytes, -v) unlimited file locks (-x) unlimited
[root@server ~]# cat /proc/cpuinfo
processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 4 model name : Intel(R) Xeon(TM) CPU 3.60GHz stepping : 10 cpu MHz : 3600.322 cache size : 2048 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 1 fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm pni monitor ds_cpl est tm2 cid cx16 xtpr bogomips : 7207.08 clflush size : 64 cache_alignment : 128 address sizes : 36 bits physical, 48 bits virtual power management:
processor : 1 vendor_id : GenuineIntel cpu family : 15 model : 4 model name : Intel(R) Xeon(TM) CPU 3.60GHz stepping : 10 cpu MHz : 3600.322 cache size : 2048 KB physical id : 3 siblings : 2 core id : 3 cpu cores : 1 fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm pni monitor ds_cpl est tm2 cid cx16 xtpr bogomips : 7199.29 clflush size : 64 cache_alignment : 128 address sizes : 36 bits physical, 48 bits virtual power management:
processor : 2 vendor_id : GenuineIntel cpu family : 15 model : 4 model name : Intel(R) Xeon(TM) CPU 3.60GHz stepping : 10 cpu MHz : 3600.322 cache size : 2048 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 1 fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm pni monitor ds_cpl est tm2 cid cx16 xtpr bogomips : 7509.03 clflush size : 64 cache_alignment : 128 address sizes : 36 bits physical, 48 bits virtual power management:
processor : 3 vendor_id : GenuineIntel cpu family : 15 model : 4 model name : Intel(R) Xeon(TM) CPU 3.60GHz stepping : 10 cpu MHz : 3600.322 cache size : 2048 KB physical id : 3 siblings : 2 core id : 3 cpu cores : 1 fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm pni monitor ds_cpl est tm2 cid cx16 xtpr bogomips : 7199.26 clflush size : 64 cache_alignment : 128 address sizes : 36 bits physical, 48 bits virtual power management:
I just want to ask. my ISP told me my server is generating high traffic from outside and paste me their traffic log with 1 IP address (xx.xx.xx.xx)
They rebooted my server and the problem disappear but I need to check what has been going on and where do I start? The only information I have is the IP xx.xx.xx.xx
I own an anime linking site which you guessed it links to anime. I have around 10000 hits a day and would like a vps with litespeed since it's so much faster.
if anyone had any recommendations for what kind of specs I would need for my website which currently recieves approx 150,000 u/v a day. I don't use any databases and all files are basically simple html files. I use approximately 15,000 GB a month (it is a flash games website).
I currently use GoDaddy's dedicated server but I fear it may not be enough; here are my current specs:
We're expecting a large spike in traffic (40k visits in one day) soon. We’re running on a very powerful server with CentOS & cPanel.
Is there any specific configuration we can setup to prepare for the large visitor spike? The website is very database and PHP intensive. We want to avoid any downtime.
We expect a site on one of our boxes to receive a significantly high level of traffic tomorrow as it is an event that will be covered and has already been covered by the press. The site operator expects > 100,000 hits a second.
It's a PHP page that pulls records from a database and lets people submit a form to insert a record. We have already put a caching script in place so that refreshing the page does not result in doing another database query.
The site was overloading a shared server, and we've moved it to one of our VPS boxes - it's the only VPS on the system at the moment. The box is a Quad Xeon 5410 with 4GB RAM with a 4 10K RPM drives in a RAID5 setup.
I have a site that has become very popular and I feel that I will soon outgrow my hosting plan. The site uses sub-domains in a fashion similar to About.com. In other words, I have different sites under the umbrella of the main domain. How does a site like About handle this issue? Forgive me if this is an uneducated question, but, since I can't host the sub-domains under different accounts and relieve some of the bandwidth burden how do I handle the bandwidth issue? It's not like I can have one account dedicated to one sub-domain and another to the other sub-domain or can I? Also, can anyone recommend a good hosting plan for popular sites? I have 60gb of transfer right now, but I want to explore other options while there is still time.
We've found out a abnormal usage of one of our servers, our RTG graphs shows:
Last 24h IfInOctets: 30.5GB MAX: 6.9MBits/s AVG: 3.4Mbits/s Cur: 4.7Mbits/s
And a strage traffic: IfOutOctets: 42.5GB MAX: 76.6MBits/s AVG: 4.7Mbits/s Cur: 600Kbits/s
We are running two websites on this server, and we looked at raw log apache, we've compilers disabled, we block most of outgoing / incoming packets on firewall, we ran chkrootkit, rkhunter and nothing was found. We checked for cronjobs, suspect files, netstat, but we can't see anything strange. We use the latest server software (apache 2.2.x), PHP 5.2.x, MySQL 4.1.x, we have most of the server optimized.
We are running iptraf now, and it seems normal: „ Incoming rates: 85.8 kbytes/sec „ Outgoing rates: 636.4 kbytes/sec Anyone have an idea? And some way to properly monitor incoming traffic? I'm looking to find how/where is the source of this traffic.
Is there any shared host which can easily drive a wordpress blog with 100,000-150,000 unique visitors every month?
The blog in question has wp-supercache and is quite a bit tweaked and consumes roughly 30-50GB of bandwidth every month (with mod deflate enabled on server). My priorities are good uptime, fast servers and network (especially to India) and good customer support.
I have a blog that gets about 50,000 unique users a month and I'm looking for a host that would fit the bill. I'm using wordpress so that's obviously a requirement, but also I would like RoR support. Other than that I'm pretty open. My fear is just that I'll get relegated to a slow server or have my account suspended. My budget is really whatever I need to pay. I'd like to find something at $15 a month or under but I'm willing to pay up to $50 or more if need be.
I'm planning on growing the traffic more in the future so I'm not sure if I should just go for dedicated hosting now or wait. I've checked out hostgator and they seem to have good reviews and fit the bill well, I'm just not sure if they are suited for high traffic sites or not.
I am writing a financial statement for my business plan and like to know what numbers should I expect from the the host. If the potential traffic will reach 405,000 visitors a month what my requirements for the hosting company should be?
I have a client with a site (wordpress blog) that gets 10,000 + hits a day. I need to find him a dedicated managed server so that his site runs smoothly and also has no outages. I just received a quote from another host for his Managed Dedicated Servers.
CPU1: Intel Xeon 5310 Clovertown (Quad Core) CPU2: Intel Xeon 5310 Clovertown (Quad Core) Total CPU Cores: Eight (8) System RAM: 6144MB (6GB) DDR2 ECC Registered System RAM Primary Hard Disk: 73GB Serial Attached SCSI (SAS) 15,000 RPM High-Performance Hard Disk Second Hard Disk: 250GB SATA-II 7,200 RPM Hard Disk (nightly backup disk) Data Transfer: 2000GB Premium Monthly Bandwidth (100Mbps uplinked port) Operating System: CentOS Enterprise Linux 5 64-Bit (x86_64) Control Panel License: cPanel / WHM + Fantastico Auto Installer$695 a month
He posts about 5-10 blogs a day too, so it's definitely a growing community website. He also has a forum with 6500 posts and 389 members.
Is this a reasonable price for a dedicated server? Would you recommend a different configuration of hardware that might make it cheaper? I would also like some examples of other sites on similar configurations if you have any, so that I can show my client what they use.
The server gets around 25k unique visitors per day, but one website in particular allows hotlinking and uses a lot of bandwidth. Last time I checked...according to whm apache status page, I was getting 180 requests per second. Not sure what time it was though. So it might be higher at a different time of the day.
Recently got mod_evasive installed, but I didn't want it to block out legitimate users. Currently it's set to this...
what a better configuration would be? When I run the log I already see it block out a bunch of IP's. I don't want to lose any visitors to this program, but I do get ddos a lot.
creating a setup that will host a site which is expected to receive 50-60K visitors in the first few hours after its launch. The site is membership based and the backend (member system) runs on PHP5-MySQL5.
Here is what I have thought of until now.
Site's sales page (which also happens to be the first page that visitors hit) hosted with Amazon S3 service. All public media files are off loaded to amazon S3 service to keep the number of requests on the hosted setup to minimum.
At the front we can have a high performance firewall like Cisco ASA 5520 followed by two dedicated load balancers in Active/Active state.
Behind the load balancers we have 3 front end servers acting as web-servers. These have SAS disks, 4GB RAM, RAID 1 setup, Dual Xeon Quad core processors each.
Behind the front end servers - we have a dedicated load balancer for the database cluster.
The database cluster consists of 3 Storage/API nodes and one of the front end servers acts as the management node. Each storage node has 8GB RAM, Dual Xeon Quad core processors, 4x RAID 10, SAS setup.
The private network is on a GigaLan.
Do you see any possible/obvious flaw in this design or anything that should be added/subtracted from the setup?
My video sharing site has high traffic, alexa rate:3,000
My site has 2 servers to split the load. 2 servers share a mysql server. Using rrdns to load the balance.
Server A running mysql 5.0,lighttpd Server B running lighttpd. Server B connect to A's mysql database.
During peak time. B can not connect to A's mysql server. It says server not responding. But A still running fine. When I check mysql log file. /usr/libexec/mysqld: Forcing close of thread .....
And when run top, the load average is 20.
The spec of Server A Intel(R) Xeon(TM) CPU 3.06GHz dual core. 2G Ram.
Here is the my.cnf
Quote:
[mysqld] datadir=/var/lib/mysql socket=/var/lib/mysql/mysql.sock # Default to using old password format for compatibility with mysql 3.x # clients (those using the mysqlclient10 compatibility package). old_passwords=1 max_connections = 1000 wait_timeout=60 connect_timeout=10 interactive_timeout=120 join_buffer_size=1M query_cache_size=128M query_cache_limit=2M max_allowed_packet=16M table_cache=1024 sort_buffer_size=2M read_buffer_size=2M
My question, do I need another maching C to run lighttpd, and just keep mysql on A. Or I can do some mysql optimization on A.
Also, if my site keeps going, can I have 1 mysql server and 5 http servers?
hosting solution for 3 very high traffic blogs, all running on WordPress.
I have been researching dedicated and I came across a couple posts where people recommended Clustered hosting over Dedicated for better handling high traffic DB driven sites in the times of Digg or Slashdot frontpage exposure.
I would like your feedback, and your opinion on what to choose from the following options:
1. Netfirms Enterprise III (Clustered)
2. ResellerZoom Failover (Clustered)
3. LiquidWeb Dedicated Webmaster Series (Dedi w/ 2GB DDR and a 3Ghz Intel Hyperthreaded)
We just ordered a new Dell server and trying to decide which flavor of Linux to use. The server is going to be used exclusively for a MySQL 5 database.
The MySQL database is very large about 9 gigs, (GIS data), and will be hit quite hard.
I was looking at CentOS and Fedora.
Here is the server specs: 2 Intel Quad Core Xeon 5410 2.33GHZ 8 Gigs Ram 2 146GB 15000RPM SAS Drives In Raid 1
