Server Stress Testing

We're having trouble with one of our servers. It has been randomly rebooting and the logs haven't indicated any problems leading up the reboots. We want to put the server through it's paces to find out if there is a hardware problem causing this.

Are there any free options which are designed specifically for load/stress testing server hardware? I'm not looking for a bash script which simply does a repetitive task. I want something that will push the server to its limits for I/O, CPU load, HD performance, etc. I've already loaded smartd for the HD but that isn't a load tester and will only tell me if the HD is slowly failing.

Any help will be appreciated.

Mathew

Random lock-ups, you gotta love 'em.

Download a boot CD with memtest and test the memory.

If the memory passes, then suspect overheating (clean dust out of CPU heatsink and make sure that it is well seated and making contact flat on the CPU die - make sure that the contact surfaces have a good smear of potting compund and isn't 'dry'.

Make sure all power connectors and add on cards are squarely seated and secure.

Look for signs of toasting around power connector blocks.

If all that looks good then suspect poor or under-rated power supply.

Luck.

list of testing tools http://ltp.sourceforge.net/tooltable.php

for free you can download yourself a copy of ultimate boot cd. this has cpu tests, memory tests etc, as well as many many other useful tools.

for free you can download yourself a copy of ultimate boot cd. this has cpu tests, memory tests etc, as well as many many other useful tools.
[/b]

I wonder why that never showed up in my searches. Since it is for work I ended up getting a license for BurnInTest Linux. Incidentally, BurnInTest didn't show up in my searches either. I ended up having to dig through our ticketing system to find what our previous Sys Eng used in the past.

glad you found a solution!!