Re: What does this mean?
On Sun, 22 Jul 2012 16:18:07 -0600, Paul E Condon wrote:
> On 20120722_132033, Camaleón wrote:
>> > Message from syslogd@gq at Jul 21 04:40:03 ...
>> > kernel:[233576.618994] EIP: [<f7de2458>]
>> > jbd2_journal_grab_journal_head+0xf/0x36 [jbd2] SS:ESP 0068:f6e83d38
>>
>> (...)
>>
>> You got a kernel oops, and Google suggests as a possible source of the
>> error a bad memory RAM stick (long mode). Being a second hand computer
>> you better run a memtest and run a pile of system stress tests to check
>> the computer components health (mainly micro, memory and hard disks).
>>
>
> I'm trying to learn. When I try to repeat your Googling the only hits
> that I get are links to *my*own* query on this list. (Not much help,
> Google. Yes I know the question.) Give me some more information on what
> query string gave you the RAM stick (long mode) answer. You've given me
> a fish. Thankyou. But I'd like to learn how to fish.
Okay, I'll show the magic but this has to be a secret bewteen us... X-)
I just searched for the line I considered suspicious (there were not many
in the logs you sent), that is, the keyword here is
"jbd2_journal_grab_journal_head" which returned a bunch of interesting
links: a bug report from openSUSE and also messages from the kernel
mailing list and both sources related to RAM as the possible cause of the
oops.
https://bugzilla.novell.com/show_bug.cgi?id=755112#c1
http://lkml.indiana.edu/hypermail/linux/kernel/1201.1/02285.html
Joining the points and given that you got a computer that was not brand
new, it makes indeed sense a hardware component problem can be hitting
you.
> Memory test and other component tests:
>
> Do you have any suggestions as to what I might download.
I prefer to run the tests from a Live system and my favorite distro for
this is SystemRescueCD but I know of others aimed to run power stress
tests to verify the harwdare component's health, such as:
http://www.inquisitor.ru/about/index.html
> I have already downloaded Knoppix v7.0.1, per Chris' suggestion, but
> have not yet found out what to do with it. Does it have memory,
> component test software on it?
I don't know Knoppix much, sorry :-( but there has to be a list of the
packages included somewhere... ah, found it:
ftp://ftp.gwdg.de/pub/linux/knoppix/packages.txt
Anyway, "memtest" is also available in Debian repos, I mean, if you
wanted to check the RAM modules, that should be enough.
> Actually, I've already learned something really significant to me: Back
> when I hit Enter to send my original post, I didn't know for sure that
> these lines were actually significant. An alternative explanation that I
> had in mind was that the kernel issued messages like these frequently
> but Gnome, or some other high level thing, trapped them and sent them to
> /dev/null. It is really nice to know that two significant people here
> take the content of these messages seriously.
Well, an kernel oops is always something to care about.
GNOME has a tool (kerneloops) that displays these messages in the user's
face and this can be nice to have because sometimes the "oopes" are
hidden in the syslog files and usually got unnoticed unless the user
experiences a serious problem with the system (reboots or hangs) and
starts raeding the logs.
> I bought the computer from a computer recycling business. They know
> about Windows, but not so much about Linux. I'd like to take some
> information to them about the nature of the problem, AND I'd like to be
> prepared to test whatever 'fix' we (the business, and I) agree to try. I
> saw this computer run Windows XP, before I took out the XP-holding disk
> and put in my own HD as the first step in installing Linux
Second hand computers are more prone to hardware failures because you
don't really know what have been they doing (power on hours) nor their
environment (there are companies that care about their machines but other
will just treat computers like fridges that open just one every 5 years
or so...). It's kinda russian roulette :-/
Greetings,
--
Camaleón
Reply to: