Task blocked for more than 120 seconds

Question

Oracle Linux 5.6 on VBox 4.3.12, on Windows 7 Pro, 64bit with 16gb of memory.
I've been picking at this problem for quite some time.  Posted some time back on the vbox.org forum, then later on the OTN Vbox forum when I discovered it.  No help there so thought I'd try attacking it from a Linux point of view.
First off --
oracle:+ASM$ uname -a
Linux vbdwprod.vbdomain 2.6.18-238.1.1.0.1.el5 #1 SMP Thu Jan 20 23:52:48 EST 2011 x86_64 x86_64 x86_64 GNU/Linux
Here's the problem.  After a VM has been up for a while it throws the following:
And the real frustration comes in that I can anticipate when it will throw this error because everything -- and I mean everything - else running on the host comes grinding to a halt.  Most other applications will start showing "Not responding".  Even if they don't show that in their title bar, they do stop responding.  Even getting Windows Task Manager up can be painfully slow, and when it does come up, it shows no stress on the CPU - still up around 95% idle.
I'm thinking some sort of clock timing issue between the guest and host, but that is really outside of my expertise.  I do add 'divisor=10' to my grub.conf as part of the initial build of all my vm's.
So, does anyone have any ideas what all this means and how I might address it?

Answer 1

Dude! wrote:
The divisor parameter helped to reduce CPU idle consumption. OL 5.6 ships with the UEK kernel as default, which is a tickless kernel, and does no longer need the "divisor=10" parameter that was used in the 5.5 and earlier versions.
What are the implications for my kernel and version?  This vm was built to be my personal test counterpart to one of my 'live' servers, so I wanted to keep kernel and version as close to that as I could.
When did you start to experience this problem? Was this working before or is this a new installation?
It's been intermittent, but experienced on 3 different host machines.  I haven't kept such close records, but if I had to guess when it started it would be at one of the upgrades to vbox.  I'll dig back through my posting history at virtualbox.org and see if I can spot anything relevant.
If it worked before, chances are that you have a hardware issue, e.g. dying disk or faulty memory. Or perhaps your virtual disk got corrupted or someting else is causing problems. Since you are on Windows, there is antivirus, defragmetnation tools, etc. that can cause very strange affect to systems inside a virtual machine.
Yeah, the first time I brought this up on virtualbox.org, they insisted it had to be a corrupt hard drive on the host.  Wouldn't even entertain the possibility of anything else.  Never mind the fact that it was occurring on two different machines at the time, never mind the fact that absolutely nothing else on either machine indicated any HD issues - including a very thorough Windows disk analysis.
Personally I have not been too happy with VirtualBox
I've had a love/hate relationship with it ever since I switched from VMWorkstation.
4.3 and found 4.2 more compatible and usable, but I think before anything else you might want upgrade to 4.3.18 and also update to OL 5.11 and see if the problem persists. Both tasks are very easy to accomplish.
Given my usage of this VM, I'd have to think long and hard about version changes on the OS.  Perhaps a point upgrade like that really isn't relevant to my concerns - I just don't know enough about that area.
Changing the version of VBox is definitely a possibility.
One other thing that might factor in somehow, and seems to relate to the link posted by Wadhah DAOUEHI, above.  When the machine is in this state and I am finally able to get Task Manger and Resource Monitor open, the one thing that seems "out there" is disk activity.  Resource Monitor will show the disk (C; drive on the host)  nearly maxed out, with over a million total b/sec attributed to vbox.exe.   That is several orders of magnitude greater than any of the other processes.
In one sense, I think I'm more concerned about why this is having such an impact on all other operations on the host.  The error as reported by the linux machine seems to be a bit of a transient, eventually getting cleared and things moving one.  But as I said, until that occurs, I can't really get anything else done either.