Bug 759837 - Raid related (possibly) failure during boot with latest koji kernel.
Summary: Raid related (possibly) failure during boot with latest koji kernel.
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel
Version: 15
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
Assignee: Kernel Maintainer List
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2011-12-04 06:55 UTC by Maciej Żenczykowski
Modified: 2012-06-07 22:24 UTC (History)
5 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2012-06-07 22:24:04 UTC
Type: ---


Attachments (Terms of Use)
Image 1 (330.02 KB, image/jpeg)
2011-12-29 12:05 UTC, Maciej Żenczykowski
no flags Details
Image 2 (458.24 KB, image/jpeg)
2011-12-29 12:34 UTC, Maciej Żenczykowski
no flags Details
Image 3 (482.10 KB, image/jpeg)
2011-12-29 12:35 UTC, Maciej Żenczykowski
no flags Details
Image 4 (476.60 KB, image/jpeg)
2011-12-29 12:35 UTC, Maciej Żenczykowski
no flags Details
Image 5 (439.66 KB, image/jpeg)
2011-12-29 12:37 UTC, Maciej Żenczykowski
no flags Details
Image 6 (482.49 KB, image/jpeg)
2011-12-29 12:37 UTC, Maciej Żenczykowski
no flags Details
Image 7 (450.56 KB, image/jpeg)
2011-12-29 12:38 UTC, Maciej Żenczykowski
no flags Details

Description Maciej Żenczykowski 2011-12-04 06:55:26 UTC
Description of problem:

2.6.40.8-4.fc15.x86_64 and earlier kernels have worked (and 2.6.40.8-4 continues to work) and boot fine.

However, 2.6.41.4-1.fc15.x86_64 fails to boot.

Images of kernel crash dumps during boot are at:
https://picasaweb.google.com/zenczykowski/FedoraKernelRaidFailure?authuser=0&feat=directlink

Version-Release number of selected component (if applicable):
2.6.41.4-1.fc15.x86_64

How reproducible:
all 10 or so boots into this kernel version have failed

Additional info, judging from the first screen shot with stack trace, the problem is probably md1_raid1 (sd_prep_fn) related.

Comment 1 Chuck Ebbert 2011-12-06 17:58:15 UTC
(In reply to comment #1)
> Description of problem:
> 
> 2.6.40.8-4.fc15.x86_64 and earlier kernels have worked (and 2.6.40.8-4
> continues to work) and boot fine.
> 
> However, 2.6.41.4-1.fc15.x86_64 fails to boot.
> 
> Images of kernel crash dumps during boot are at:
> https://picasaweb.google.com/zenczykowski/FedoraKernelRaidFailure?authuser=0&feat=directlink
> 

Please attach the images to the bug report.

Comment 2 Maciej Żenczykowski 2011-12-29 12:05:08 UTC
Created attachment 549936 [details]
Image 1

Comment 3 Maciej Żenczykowski 2011-12-29 12:34:52 UTC
Created attachment 549947 [details]
Image 2

Comment 4 Maciej Żenczykowski 2011-12-29 12:35:26 UTC
Created attachment 549948 [details]
Image 3

Comment 5 Maciej Żenczykowski 2011-12-29 12:35:59 UTC
Created attachment 549949 [details]
Image 4

Comment 6 Maciej Żenczykowski 2011-12-29 12:37:02 UTC
Created attachment 549950 [details]
Image 5

Comment 7 Maciej Żenczykowski 2011-12-29 12:37:45 UTC
Created attachment 549951 [details]
Image 6

Comment 8 Maciej Żenczykowski 2011-12-29 12:38:21 UTC
Created attachment 549952 [details]
Image 7

Comment 9 Maciej Żenczykowski 2011-12-29 12:47:46 UTC
As requested I've attached the images to the bug report (terrible upload interface).

---

As an additional note, I've now got another machine that exhibits similar symptoms: it boots with 2.6.40 and fails to boot with 2.6.41.

I'm guessing it's the same failure mode, although once again, I can't actually log the messages and they scrollback very quickly on the 80x25 terminal.

This new configuration is significantly simpler - no LUKS and no LVM.  Simply software raid.
sda/sdb are 2.5" SATA drives, sdc/sdd are USB mass storage devices.

Here's what /prod/mdstat looks like on 2.6.40.8-4

# cat /proc/mdstat 
Personalities : [raid1] [raid0] 
md4 : active raid0 sda4[0] sdb4[1]
      929536768 blocks 64k chunks
      
md3 : active raid1 sda3[0](W) sdb3[1](W)
      15791544 blocks super 1.0 [3/2] [UU_]
      bitmap: 7/8 pages [28KB], 1024KB chunk

md2 : active raid0 sdc1[0] sdd1[1]
      15791616 blocks 64k chunks
      
md1 : active raid1 sda2[0](W) sdb2[1](W)
      7550536 blocks super 1.0 [3/2] [UU_]
      bitmap: 8/8 pages [32KB], 512KB chunk

md0 : active raid1 sda1[0](W) sdb1[1](W)
      273060 blocks super 1.0 [3/2] [UU_]
      bitmap: 1/2 pages [4KB], 128KB chunk

unused devices: <none>

As you can see, just raid0/1 and not super complex.  /dev/md{0,1,3,4} are ext{3,4,4,4} file systems.

Comment 10 Josh Boyer 2012-06-07 15:15:29 UTC
Did this resolve itself with 2.6.43/3.3?

Comment 11 Maciej Żenczykowski 2012-06-07 22:24:04 UTC
Yes, at some point around 2.6.42/43 3.2/3.3 this resolved itself.
Closing the bug.


Note You need to log in before you can comment on or make changes to this bug.