PHP newbie - scared to death!

[dbookpg]

Newbie here...first post. I'm super nervous and hoping to get some advice? I am a professional developer and have worked for 10 years in the microsoft world. I make my living managing several online applications catering to different types of sales people. Generally, CRM (contact management) type stuff. Anyway, 100% of the products I "charge money for" are ASP, ASP.net and all are exclusivley MS SQL.

 

So... why the PHP board? About a year ago I decided to take the plunge into PHP/mySQL - MS piss### me off!. I'm super stoked. I have re-created one of my main "livings" and plan to go commercial with it in about a week? I have one very large customer that will be the only customer using this app. It's php/mySql and it tests to my satisfaction. My concern...not sure. I have this application running on a dedicated box and everything seems ok. I'm looking for some input as to what types of "things" I should be monitoring over the first few weeks/months of live use. There will be around 300 users, generally about 30 logged in at any given time. Are there any monitoring tools or whatever for this strange UNIX thing? I'm sooooooo used to using my windows PC to "Check up" on miscellaneous servers, etc. I have no idea how I'm going to accomplish this with this UNIX rig. The box is a dedicated server located at an ISP. I have been happy with the ISP's service since "renting" the box about a year ago (during development).

 

Just looking for some insights so I can sleep at night when it's "live"

 

thanks in advance.

David.

[btherl]

Here are some of the common tools I use

 

top - displays processes using the most processor power.  Also gives a summary at the top of "system time" (usually disk reads/writes)

ps - displays detailed information about processes

iostat - displays details about i/o

tail -f /var/log/apache/access.log - Shows requests as they come in (if your logs are in the default location).  May not be useful if you have 30 concurrent users unless you filter the data with "grep".

grep - Possibly the most useful tool for unix.  It displays lines matching a particular expression, perfect for extracting information from log files

 

Those are all quite crude methods, and won't give you much detail.  But they will tell you if your box is about to blow up due to lack of some critical resource.  They are a good first stop for diagnosing problems.

 

The other important number is the load (displayed by top and w).  Load indicates the number of processes waiting for something.  If it's high, then you probably have a problem.  Through observation you can tell when system performance is impacted by a high load figure (this depends very much on your system), and then setup an alert for that load figure.

 

Oh, a good monitoring package is nagios.  It can be configured to alert you when things are exceeding limits or not functioning.  Good things to monitor are load, cpu usage, disk space, and response time for http requests.

[pyrodude]

For the first month or so, it wouldn't hurt to do a daily backup.  After that, (and depending on the sensetivity/necessity of the data) you may want to continue wit daily backups or perhaps switch to weekly.  Should something happen to your lovely application, you wouldn't want your client to be SOL!