HP Workstation Hangs Randomly

T

TobyJamesJoy

Hi,

I'm an extremely green *BSD user trying to run FreeBSD 9.1 on my HP xw4600 workstation. After numerous install attempts using several different install mediums (boot only CD, USB .img, PC-BSD USB .img) and different install methods (FreeBSD handbook method, RootOnZFS GPT and MBR methods from the wiki, PC-BSD handbook method) I keep running into system hangs. These occur regardless of whether I'm logged in over the network or not.

When the system locks up the console cursor doesn't flash and I can't change the tty, manually powering off is the only way back in. I've tried to reproduce the issue with an Ubuntu live CD as well as an Arch Linux install with no success.

Is there some important configuration step I'm missing? I've searched the web, the hardware compatibility list and these forums but found nothing useful, sorry if I've missed something obvious.

Thanks for your time,
Toby.
 
Hi John, yes I have, there only thing that stands out is this line:

Code: 
kernel: pid 79165 (try), uid 0: exited on signal 10 (core dumped)

That appears about three hours before the final message for that boot though.
 
That message is usually from a configure run when building ports. Do you use pre-build packages or did you build any ports prior to the crash?
 
In that case the message is not related to the crash. Which version of FreeBSD, amd64 or i386? How much memory does the machine have and is there any swap configured?
 
Is the system using ZFS and ZFS on root install? You may be encountering a common gotcha with ZFS that is the ARC cache consuming too much wired memory (memory that can not be swapped). If that is the case put this into /boot/loader.conf and reboot:

Code: 
vfs.zfs.arc_max="1024M"
 
This current install is just UFS but on other ZFS installs I tried the problem did seem to occur more frequently. Is there a UFS equivalent for that line?
 
This problem keeps occurring, at random intervals too. I'm going to try again on some different hardware when I can get my hands on it, until then I'll (begrudgingly) go back to Linux. Thanks for the help.

Toby.
 
Run a RAM memory test. I have seen this type of behaviour on a machine with bad memory. I have also seen good memory go bad. So good idea just to rule it out, just in case.
 
I would like to support dave. Some months ago I also experienced random freezes when setting up a new machine. I replaced the RAM and since then I had no similar failures anymore.

Some observations. This happened when summertime started here in Brazil, and at night at lower ambient temperature it took longer until the machine froze, so I was able to do some basic setups. Compiling huge ports froze the machine in any case. It would have been a nice experiment to put the CPU into the fridge, but unfortunately (for the science) the new memory arrived, and that resolved the problem.

It is hot in Melbourne right now, isn't it? :-)


PS: Other flaky hardware may cause this too.
 
Thanks dave and rolheinrich, I've already wiped FreeBSD from my drive and installed Arch Linux, but under Arch I've run a couple of RAM tests and nothing has showed up. It's probably some other component but so long as things continue to run under Linux I'll use it as a desktop and search for more reliable server hardware to get my FreeBSD on.

Thanks for your help.
 
rolfheinrich said:
I would like to support dave. Some months ago I also experienced random freezes when setting up a new machine. I replaced the RAM and since then I had no similar failures anymore.

Some observations. This happened when summertime started here in Brazil, and at night at lower ambient temperature it took longer until the machine froze, so I was able to do some basic setups. Compiling huge ports froze the machine in any case. It would have been a nice experiment to put the CPU into the fridge, but unfortunately (for the science) the new memory arrived, and that resolved the problem.

It is hot in Melbourne right now, isn't it? :-)


PS: Other flaky hardware may cause this too.
Click to expand...

It certainly has been hot in Melbourne recently but I don't think this is from stressing the system, the freezes are just as likely to occur when the machine is idle as when compiling.
 
You must log in or register to reply here.
Back
Top