Thread: How would you do health check of Linux/UNIX servers?

Hey.. Greetings to All...

I am a new user to this forum.. nixcraft seems to be quite interesting..

My first question here is - As a System Administrator, what are those checks which would ensure that our Linux/UNIX servers are in healthy state; besides checking CPU,memory,load average utilizations?

Thank you.

Health check is a broad topic. It all depends upon your server role. If server is a load balancer than; I'd monitor LB software and network closely apart from system load. Try

1. Logwatch - Watch your logs in /var/log/* for errors, security and warnings and sends daily email summery report about them.
2. Nagios - a popular system and network monitoring software application. It watches hosts and services, alerting users when things go wrong and again when they get better including cpu, memory, network, log files, disks, MS-Windows clients, server temperature & over all health and tons of other things.
3. sar - It collects, report, or save system activity information last 30 days or so. It can answer why load was high at certain time of the day.
4. There are many other software and plugins to monitor specific part of the system. If you need any specific info let us know.

Okay, let me elaborate more.
Suppose there are three web-sites which are hosted on apache and tomcat servers. I am the SA and I want to check if my server is responding to customer's requests properly. Besides using automated monitoring tools (nagios,sitescope,splunk), what would be the steps which I can perform manually on my servers to ensure servers are in healthy condition?

2. Second question.. Let me take a scenario here.. For example yahoo.com is working at my location but a user at xyz location cann't access it. Say Yahoo.com is hosted on apache web servers, what would be the basic troubleshooting steps which SA will execute from the server end?

Thank you and let me know if its not enough information for you to answer.

No one replied still..??

Dav86,

It is weekend so things get slower around here. Anyway, the number one rule for troubleshooting UNIX / Linux problems is to go though log files. If I were you I will start as follows:

• Check Apache log files
• Check Apache error log files
• Make sure apache is running on port 80
• Make sure firewall is not blocking port 80
• Make sure disk i/o and memory is under control (top, free, vmstat, sar and there are tools to monitor this)
• Use ps to find out exact memory consumption for apache and again make sure it is under control
• If particular user is not able to access a website, get his ip and compare it with your firewall and apache log files.

These are basic steps. If you are really new to Linux, learn bash and purchase good books which teaches basic commands and troubleshooting. It takes time to master everything, have patience

Good luck!

Hey - This is helpful. Thanks a lot..!!



I am not completely new, even I work on RHEL5 production servers though at junior level – I have root access but cannot do research on those ; it might hamper customers...
It is like I never tried to get into depth of fundamental concepts of Linux OS. I am trying to get clarity on important topics, yeaa.. I know, need to be patient, becoming expert is not that easy….