Dr. Sol Adoni
Podcasts - SOLLOG - Thee End - Real Paranormal
One of the major things that kills server load are scripts that suck up huge VAR useage.
Most server ADMINS are clueless as to how to properly manage VAR.
When you set up a server you should allocate substantial allocations to /VAR
The reason is, cPanel runs in /VAR which is used now on most servers that run a Linux OS.
PHP files for scripts such as WP WordPress and Forum scripts like PHPbbs also suck up lots of space in VAR.
To see your VAR useage use these commands via SSH with ROOT access.
du -h /var/ –max-depth=1
du -h /var/lib/mysql/ –max-depth=1
There’s nothing you can do if you have heavy traffic and use PHP database scripts like WP or a busy PHP Forum, so load up on your VAR allocation as to how much disk space you give to it when you set up a server you expect to be busy.
If you have a 1TB HD you should have a mirror 2nd drive of same size and use a RAID to copy all essential files in case your primary HD and it’s VAR partition fail. Then you should do at least daily if not 2X Daily remote backup of your data incase your NOC has a major problem like a fire or something.
BACKUP is seldomly done properly by most companies.
Heavy use of database scripts such as WP or phpLIST (an email management script) and Forums such as PHPbbs will suck up lots of VAR use.
10% to even 25% of HD should be allocated to VAR in my opinion due to the nature of heavy use of VAR by most common scripts today, so a 1TB HD means 100G to 200G in VAR allocation. That seems like a lot, but if you run busy sites using WP or a BBS, you will use lots of VAR.
Then the more MEMORY you have on the server, the better VAR will perform, so 32G and even 64G of memory for VAR intensive servers would be a great idea.
Next to the speed of the CPU and how many CPU’s you use, large memory and fast drives are essential for server load issues.
A well setup server should not run more than a 1% CPU useage and heavy use of VAR intensive scripts will bog down a server’s load handling abilities.
Today 1G is a decent uplink to the NOC, but 100MPS was the standard for a fast server for years.
10TB to 30TB bandwidth are standard allocations today for bandwidth useage on busy servers.
MySQL is a standard SQL program many networks now use to do basic database management, yet very few ‘admin’s really understand how to properly manage heavy MySQL or any SQL allocations and how to track down server load issues caused by such heavy SQL useage.
SQL is just Structured Query Language and it’s not new, the main paper for SQL protocols was written way back in the era of mainframes in 1970 or so.
So if you are having heavy server load useage issues or getting /VAR notices due to /VAR being nearly maxed out, use the commands above to find out what is hogging up VAR on your server.