Bug 179729
Summary: | Slab corruption in Devel Kernel | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Daniel Walsh <dwalsh> |
Component: | kernel | Assignee: | Dave Jones <davej> |
Status: | CLOSED RAWHIDE | QA Contact: | Brian Brock <bbrock> |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | rawhide | CC: | pfrields, pinter, wtogami |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2006-08-02 02:01:27 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Attachments: |
Description
Daniel Walsh
2006-02-02 14:56:59 UTC
Created attachment 124049 [details]
Slab.txt
Created attachment 125283 [details]
logfile segment showing kernel oops, then slab corruption detail
Not sure if what I'm seeing here is germane, but I'm freezing regularly on a
PIII 600MHz 512MB(ECC) with no discernable usage pattern to trigger it.
I've swapped memory sticks in place, moved IDE from ATA33 onboard to Maxtor
ATA133 PCI card, installed an Antec power supply and swapped out the DVDROM
drive with an LG-4167B trying to pin this down.
All FC5 kernels through 2.6.15-1.1977_FC5 exhibit the system hang symptom,
although it isn't always logged. When it is logged, I save the detail away.
I've attached a section of a recent log showing both a kernel oops ('Unable to
handle kernel paging request') followed shortly thereafter by a 'Slab
corruption: (Not tainted) start=d2f73130'.
Please let me know if there is anything I can do to help isolate this further.
Cheers,
/p2
Has this reoccured recently ? There have been a number of slab corruption problems fixed recently. Last freeze was March 1st, on kernel build 1977 (no log detail available). I updated to build 1996 at that time, and so far everything is stable. Here's hoping... Created attachment 125951 [details]
logfile segment showing 'Drive appears confused', then slab corruption
Okay, I may have been dealing with two problems.
In the large, slab corruption issues seem to have cleared insofar as system
freezes have dropped markedly. When it does hang, and I get some log detail, it
appears related to IDE handling of my dvdrom.
As previously noted, I've replaced both the controller and the dvdrom unit
recently, but still no joy. I thought keeping a disc in the drive seemed to
help, but I'm not sure now.
When the machine hangs, comes back, then locks up hard all within a minute or
two, I usually get some log detail. Otherwise, it just freezes and I get
schmaltz.
I'm upgrading to kernel build 2038 now. Once booted, I'll attach the boot log
showing the irq map, if that helps. Otherwise let me know if there's anything
else needed.
Thanks!
Created attachment 125979 [details]
logfile segment showing 'Drive appears confused', *no* slab corruption
Some good news, I think.
Since upgrading to kernel build 2038 and stress testing the dvdrom burner for
~8 hours in every way possible locally, I cannot induce a freeze at all.
As the attached log shows, I still get those knarly 'Drive appears confused'
warnings, but no other slab corruption errors and the machine just hums along.
Cheers,
/p2
Regrettably, since that last post a week ago two system hangs have occurred that, when logged, show the fargin 'Drive appears confused' warning as the last gasp text recorded before hard lock. I thought maybe it was VMware Server Beta build 22088, but it wasn't running today at all. Only a NFS read of a 16GB vmware image to my backup server on OS X. Gnome was running too, if it matters. I'm on 2.6.15-1.2054_FC5. Have been running stable on a 2.6.17 kernel for a while now and although I still get fist fulls of those 'drive appears confused' messages, the hard freezing appears fixed despite all my attempts to stress the box. |