On 05/04/2011 04:36 AM, PsychoGTI wrote:
>
> I am by no means a “Flux Capacitor Penguin”
i guess you know that that means absolutely nothing, but one thing: has
posted a lot, needs to “get a life” etc…
says nothing about skill or knowledge…i wish they would get rid of
it…i tried my best to avoid it for over a year by posting using a
series of unregistered nom de guerre, but strangely some of the mods and
administrators here thought i was up to no good…and, after a direct
threat of banning i gave in and let the counting begin…
> Is there a way to test the write to swap ability?
normally the linux kernel will stuff some things into swap even before
RAM is full…i have zero idea how it goes about deciding what to put
where and when…but, top just now tells me i have been up 65 minutes
and have 3168k in swap with ram near full and about 43% of that cached…
so, it is because you wrote “This whole time the swap remains
untouched.” that i suspect that might be what cause the kernel to
complain “Marking nosave pages”
hmmmm…i googled “Marking nosave pages” and found 21k hits
with problems associated like acpi, hibernate, display and after
scanning a few pages of hits, i changed the search string, and leave it
to you to dig into these ~65 deeply: http://tinyurl.com/6f6mh5j
>> and copy paste that to paste.opensuse.org
>
>
> Will do. I’ll type out a short description of the problem, and then
> paste in the /var/log/messages from the start of the scan to the time is
> dies (at the Marking No Save Pages).
>
sorry, i see i failed to mention to paste the URL to the paste page back
to this thread (which is the only way we have access to it)
> I would really like to run the OS drives and DATA drives in RAID.
i’m not suggesting you change your production machine, just looking for
a way to prove or disprove it is a drive problem (most likely a RAID
problem) by eliminating those from the mix during a test…
> [big snip of RAID output]… not sure what that
> means, and if it is part of the problem.
>
> Do you know of any log or config I can check to see what the overall
> /dev config should be for mdadm devices?
as mentioned, when it comes to RAID i’m an idiot…except to know that:
-software raid is junk (invented by folks who $ELL $oftware)
-no RAID is a good substitute for a proper backup routine
-only good if a 24x7x365 hot replacement is required
-almost always is more difficult to setup/administer than none
so, as a simple man i’ve not bothered to learn much about it…
if you have a 24x7x365 commercial commitment then you are [imnsho] in
the wrong place to begin with…openSUSE is a short life, near cutting
edge, consumer level distro where lots of enthusiast discover LOTS of
bugs…which get worked out and eventually the code becomes clean
enough for Novell to release the commercial applications ready SUSE
Linux Enterprise Server (SLES) or SUSE Linux Enterprise Desktop (SLED)…
i have zero idea, but it might be that their latest is already set to
work and play nicely with your setup…
anyway, once you have nailed down where the bug is, i will ask you to
log a bug so that it can be fixed…
>
> I called it “Kernel Failure” because I can find no other indications
not sure, but i think Dave’s point is that an uncommanded reboot might
just be a momentary loss of ground in a hardware switch…while most
“kernel failure” are total system hard freeze, kernel panic, etc…
in fact, i think it really good to listen to him (he is a real guru)
and his post seems to really be saying: it is not a kernel problem, a
declaration i can neither prove or disprove…however, i think the
single hard drive test will prove your system stable without RAID.
>
> Any other logs and such I should be checking besides
> /var/logs/messages?
well, you could look for the strings “error” and/or “warning” in *:
/var/log/boot.msg
/var/log/warn
/var/log/Xorg.0.log
hmmmm, that last one caused me to think…if you wish, before you do
surgery (removing drives and replacing with one, for a test) you might
want to remove X from the equation…you have not declared which AVG
you are running but i’m pretty confident whatever it is is doesn’t
require a GUI, so with the system and hardware you have boot to runlevel
three and see if it will complete the scan with rebooting…if it does
maybe it is an X problem, and neither a kernel nor RAID problem!
–
CAVEAT: http://is.gd/bpoMD
[openSUSE 11.3 + KDE4.5.5 + Thunderbird3.1.8 via NNTP]
HACK Everything → http://www.youtube.com/watch?v=j5b4CCe9pS8&NR=1
*