[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] xen-4.4.0 and unexpected system reboot

"squidmobile@xxxxxxxxxxx" <squidmobile@xxxxxxxxxxx> writes:

> 26 jun 2014
> greetings,
>> It might be a good idea to turn off the screen saver for the console if
>> you do this.  If the crash happened after the screen saver became
>> active, you might see nothing but a blank screen.  Guess how I know
>> ... :)
> *rofl'ing*  nothing like a little first hand experience to make an
> impression, no?

Well, I had the camera ready to take a screenshot to finally get a hold
of the messages, so it was a bit disappointing that the screen was only
black ...

It's kinda embarrassing when you start doing things like that.

>> I'm finding the messages about loosing contact to the UPS
>> suspicious.  I used nut quite a while with an USB connection, and no
>> matter what load I put on the CPU, I never noticed any such messages.
>> Do you see a correlation between CPU load and these messages?
> i saw another message the other day, and it was the one about low
> battery.  since the battery is only a few months old, i think
> something is odd with the ups in that context.  it claimed a
> brand-new 12-volt battery provided only about 4.9 volts, which is
> what it said about the old one i replaced.

Perhaps the new battery is bad.

> my ups also connects via usb.  i forget the exact verbage from the
> nut messages, as i've seen several of them.  i figure some of them
> are related to ntp adjusting the system clock, which could make nut
> think it missed its handshake.

Hm, I don't remember if you can set a polling interval with nut.  You
can with apcupsd, and I've set it to a minute or so.  Clock changes
through ntp shouldn't be able to mess up such long time intervals.

> yes, i see a correlation between cpu load and nut messages.  i use
> software raid1 for two disks with encrypted partitions on my disks.
> plus, lvm and jfs.  if i do something extremely disk i/o intensive,
> it drives my cpu loading through the roof.  and nut starts sending
> out broadcast messages, which doesn't help my train of thought.

Increase the polling interval?  You probably don't want nut to perform a
shutdown because of heavy disk I/O :)

>> If there appears
>> to be any sort of regularity to the crashes (like every 45 days), I'd
>> try to find out what is being started every 45 days.
> no regularity that i've seen so far.  no automated tasks that
> suddenly kicked in, although i'm sure i ran <something> at the
> time.  i don't remember just what.

Hm, just wait and see if happens again ...

Knowledge is volatile and fluid.  Software is power.

Xen-users mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.