Server hangs or freezes can occur for a number of reasons, but I find there are just a few causes. Sometimes it may just be a single issue that causes a hang, but often the server freezing up or crashing is caused by a number of problems happening simultaneously.
Here are a few of the most frequent issues I see that causes a server to hang.
Network Problems
Local, regional or at the data center, rackAID receives reports every week of servers being down or frozen. Typically, clients say they cannot get to the website or get a web site not found error. In most cases, the server is fine, but there is a network problem.
I use Alertra’s Spotcheck to verify a site is up. Spotcheck runs from numerous locations so you can tell if your site is really down or you have a network issue.. If more detail is needed, then you may need to learn how to run a traceroute.
System Overload
Server hang-ups due to system overload typically show up as slowly loading pages. The system may respond slowly for a period and then return to normal operation. Server slow downs typically happen at peak traffic periods, when scheduled tasks (cron jobs) run, or when certain maintenance occurs. Keeping track of the server freezes can help pinpoint what maybe happening.
To help with server overloads, I use sysstat. We include this tool on all server management services we provide. Sysstat keeps track of performance metrics which can make troubleshooting system freezes easier. The historical data can be cross referenced with the logs you keep about server freezes to better pinpoint the cause.
Configuration Errors
Nearly weekly, someone contacts us with a hung server that is offline due to mis-configuration. If you use the command line, an errant quotation can bring Apache, Sendmail or other critical systems to a halt.
To quickly fix configuration issues, take backups. This way you can roll back to a prior configuration file. Or simply, stay out of the command line and use a tool like Plesk or cPanel to configure your system.
Hardware Issues
I love how people like to blame the hardware. But in my experience, hardware is rearely the cause. Typically, if there are hardware problems, you will see random freezes and hangs from day one. Otherwise, the system will run great for years.
The only exception to this is disk drives. Disk drives fail and you never know when it is going to happen. So take backups and be prepared.
Backups
I know. I just said you need backups, but in many cases, I’ve seen backups cause servers to hang. For example, rsync, used on cPanel and many custom scripts, consumes a significant amount of CPU resources which can slow your server and even cause it to freeze up. So while you need backups, you don’t want them hanging your server.
I recommend using something like our Linux server backup solution. Our R1Soft based server backup service can backup very busy systems with minimal impact to performance.
Your Reasons?
If you have more tips about server freezes, let me know. I find these to be some of the key culprits and the first place to look if your server is hanging frequently.