Scratching my head here. Hoping someone can help me troubleshoot.

I have a Dell PowerEdge SC1435 server which had been running with a previous version of ubuntu for a while. (I believe it was 13.10 server x64)

I recently reformatted the drive (SSD) and installed ubuntu server 14.04.1 x64.

All seemed fine through the install but the machine hung on first boot at the end of the kernel output, just before I would expect the screen to clear and a logon prompt appear. There were no obvious errors at the end of the kernel output that I saw. (There was a message about "not using cpu thermal sensor that is unreliable" but that appears to be there regardless of whether it boots or not)

I gave it a good 5 minutes and then forced a reboot. To my surprise it booted to the logon prompt in about 1-2 seconds after bios post. I rebooted again and it seemed to pause for a few extra seconds where it hung before, but proceeded to the login screen. Rebooting again it was fast again. So at this point I thought it was just one of those random one-off glitches that I would never explain so I moved on.

I installed a few packages (exact same packages installed on the same OS version on other hardware), did apt upgrade and dist-upgrade then rebooted. It seemed to hang again so I drove to the datacentre and connected a console only to get a blank screen. Forced reboot again. (also setup ipmi for remote rebooting and got rid of the grub recordfail so it would not wait for me to press enter!)

That was very late last night. I came home, did a few reboots with no issue so went to bed. Today I did a reboot again to check it and again it crashed somewhere. I remotely force rebooted it.

As this point I started digging a little more and immediately noticed something really strange.

top - 14:18:35 up 8 min,  1 user,  load average: 1.00, 0.85, 0.45
Tasks: 148 total,   1 running, 147 sleeping,   0 stopped,   0 zombie
%Cpu(s):  0.1 us,  0.3 sy,  0.0 ni, 99.6 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:  33013620 total,   338928 used, 32674692 free,     9740 buffers
KiB Swap:  3906556 total,        0 used,  3906556 free.    47780 cached Mem

    1 root      20   0   33508   2772   1404 S   0.0  0.0   0:03.82 init
    2 root      20   0       0      0      0 S   0.0  0.0   0:00.00 kthreadd
    3 root      20   0       0      0      0 S   0.0  0.0   0:00.00 ksoftirqd/0
    5 root       0 -20       0      0      0 S   0.0  0.0   0:00.00 kworker/0:0H
    6 root      20   0       0      0      0 S   0.0  0.0   0:00.00 kworker/u16:0
    8 root      20   0       0      0      0 S   0.0  0.0   0:00.24 rcu_sched
    9 root      20   0       0      0      0 S   0.0  0.0   0:00.02 rcuos/0
   10 root      20   0       0      0      0 S   0.0  0.0   0:00.00 rcuos/1
   11 root      20   0       0      0      0 S   0.0  0.0   0:00.00 rcuos/2

This server is completely unused and idle, yet it has a 1 minute load average of exactly 1.00? As I watch the other values - the 5 minute and 15 minute also appear to be heading towards 1.00 so I assume they will all reach 1.00 at some point. (The "1 Running" is the top process)

I have never had this before and since I have no idea what is causing the startup crashing, I am assuming at this point that the two are likely related.

What I would like to do is identify (and hopefully eliminate) what is causing that false load average and my crashing issue.

So far I have been unable to identify what process could be waiting for a resource of some kind to generate that load average.

I would very much appreciate it if someone could help me to try and track it down.

top shows all processes pretty much always sleeping. Some occasionally popping up top but I think that's pretty normal. CPU usage is mostly showing 100% IDLE, with very occasional dips to 99% or so.

nmon doesn't show me much. everything just looks idle.

iotop shows pretty much no traffic whatsoever. (again, very occasional spots of disk access)

interrupt frequency seems low. way below 100/sec from what I can see.

I saw numerous google discussions suggesting this:

echo 100 > /sys/module/ipmi_si/parameters/kipmid_max_busy_us

..no effect.

RAM in the server is ECC and test passes.

Server install was 'minimal' (F4 option) with OpenSSH server ticked during install. Installed a few packages afterwards including vim, bcache-tools, bridge-utils, qemu, software-properties-common, open-iscsi, qemu-kvm, cpu-checker, socat, ntp and nodejs. (Think that is about it)

I have tried disabling and removing the bcache kernel module. no effect. stopped iscsi service.. no effect. (although there is absolutely nothing configured on this server yet)

I will leave it there before this gets insanely long. If anyone could help me try to figure this out it would be very much appreciated.





Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.