New subject: [Bug 601328] System looses 3ware (or qlogic hda) PCIe devices after up to 4 weeks of runtime.

30 Apr 2010

      http://bugzilla.novell.com/show_bug.cgi?id=601328

http://bugzilla.novell.com/show_bug.cgi?id=601328#c0

           Summary: System looses 3ware (or qlogic hda) PCIe devices after
                    up to 4 weeks of runtime.
    Classification: openSUSE
           Product: openSUSE 11.2
           Version: Final
          Platform: x86-64
        OS/Version: openSUSE 11.2
            Status: NEW
          Severity: Critical
          Priority: P5 - None
         Component: Xen
        AssignedTo: jdouglas@novell.com
        ReportedBy: claus@soonr.com
         QAContact: qa@suse.de
          Found By: ---
           Blocker: ---

User-Agent:       Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.8)
Gecko/20100204 SUSE/3.5.8-0.1.1 Firefox/3.5.8

Our system uses Intel E5530 on SuperMicro X8DTU-F motherboard, has 3ware 9690SA
controller and a qlogic 2460 HBA. Under Linux xen11 2.6.31.12-0.2-xen (and
previous) kernel we would see that the PCIe devices would "disappear" with
messages like "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this
message.

Same HW platform seems more stable under Linux xen10 2.6.31.12-0.2-desktop #1
SMP PREEMPT. 

There is something odd with Xen kernel message timing - dmesg reports events
with out of order timestamps - e.g.

[437054.738244] XFS mounting filesystem dm-5
[437054.833957] Ending clean XFS mount for filesystem: dm-5
[437075.264982] device vif2.0 entered promiscuous mode
[435331.784715] br3: port 3(vif2.0) entering forwarding state
[437047.598690] physdev match: using --physdev-out in the OUTPUT, FORWARD and
POSTROUTING chains for non-bridged traffic is not supported anymore.
[437075.289243] physdev match: using --physdev-out in the OUTPUT, FORWARD and
POSTROUTING chains for non-bridged traffic is not supported anymore.
[437075.289256] physdev match: using --physdev-out in the OUTPUT, FORWARD and
POSTROUTING chains for non-bridged traffic is not supported anymore.
[435331.818121] (cdrom_add_media_watch()
file=/usr/src/packages/BUILD/kernel-xen-2.6.31.12/linux-2.6.31/drivers/xen/blkback/cdrom.c,
line=108) nodename:backend/vbd/2/832

Also - under normal kernel "dmesg | grep micro" reports (16 times - 8 cores
with hypertrhreading):

[    9.364947] microcode: CPU0 sig=0x106a5, pf=0x1, revision=0x11
[    9.364954] platform microcode: firmware: requesting intel-ucode/06-1a-05

BUT under Xen kernel - ONLY ONE line:

[   11.797309] platform microcode: firmware: requesting intel-ucode/06-1a-05

Perhaps platform is unstable due to timing issues under Xen ?

Reproducible: Always

Steps to Reproduce:
1.Boot HW under Xen kernel.
2.Have load on fiber HBA and 3ware controller.
3.Wait <= 2 weeks for crash (sometimes within 24 hours).

See the note above regarding "dmesg | grep microcode" aqnd out of order
timestamps for dmesg (UNDER Xen).

-- 
Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug 601328] New: System looses 3ware (or qlogic hda) PCIe devices after up to 4 weeks of runtime.

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

bugzilla_noreply＠novell.com

tags

participants (1)