Bug 612036

Summary: Lot of "btrfs bad tree block start" Errors
Product: [Fedora] Fedora Reporter: Michael Hagmann <michael.hagmann>
Component: btrfs-progsAssignee: Josef Bacik <josef>
Status: CLOSED DUPLICATE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: high Docs Contact:
Priority: low    
Version: 13CC: jbacik, josef, mmahut
Target Milestone: ---Keywords: Reopened
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2010-08-29 16:34:10 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 Flags
sosreport KVM Host with the errors
none
sosreport KVM Host with the errors ( all options enabled ) none

Description Michael Hagmann 2010-07-07 05:53:25 UTC
Created attachment 429963 [details]
sosreport KVM Host with the errors

Description of problem:

After a fresh installation with some kvm guests on Fedora 13 and one Debian Lenny

I copyed in the Fedora13 Guest on a separate xfs lvm-volume handover to the guests, then the errors starts 

Version-Release number of selected component (if applicable):

see sosreport

How reproducible:

?

Steps to Reproduce:
1.
2.
3.
  
Actual results:

KVM Guests are frozen

Expected results:

non


Additional info:

Comment 1 Michael Hagmann 2010-07-07 05:58:31 UTC
Created attachment 429965 [details]
sosreport KVM Host with the errors ( all options enabled )

sosreport KVM Host with the errors ( all options enabled )

Comment 2 Josef Bacik 2010-07-07 12:55:22 UTC
Jul  5 00:57:08 enif kernel: sd 6:0:1:0: [sdb] Device not ready
Jul  5 00:57:08 enif kernel: sd 6:0:1:0: [sdb] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
Jul  5 00:57:08 enif kernel: sd 6:0:1:0: [sdb] Sense Key : Not Ready [current]
Jul  5 00:57:08 enif kernel: sd 6:0:1:0: [sdb] Add. Sense: Logical unit not ready, cause not reportable
Jul  5 00:57:08 enif kernel: sd 6:0:1:0: [sdb] CDB: Write(10): 2a 00 3e 86 31 ff 00 00 18 00
Jul  5 00:57:08 enif kernel: end_request: I/O error, dev sdb, sector 1048982015

and then it looks like you had an unclean shutdown.  This is the partition with your kvm vm's on it I assume, the fs was corrupted.  You'll have to reformat.

Comment 3 Michael Hagmann 2010-07-07 19:53:18 UTC
Hi Josef

one question, why the FS don't handle that kind of error ? Is that development not yet done ?

When the problem occured I wrote with a rsync on a xfs FS that had not a such severe Problem. I could fix it with xfsrepair.

thanks for your explanation. 
Mike

Comment 4 Josef Bacik 2010-07-07 20:07:48 UTC
Yeah basically.  Alot of the error handling code hasn't been completed (we just do BUG_ON() if we run into an error) or it hasn't been tested very well.  We're more focusing on features and stability.  It is definitely something that needs to be worked on, but there are a lot of parts of btrfs that are in that state :).

Comment 5 Michael Hagmann 2010-07-07 20:28:21 UTC
Josef

could you use my corrupted diskimage ? or could I provide more Information before I wipe out the date ?

thanks Mike

Comment 6 Josef Bacik 2010-07-07 20:48:19 UTC
Mike,

I can't do anything with it, but Yan may need it.  Go ahead and use btrfs-image to save an image, it has a compression option that you can use, and it doesn't save the data, it just saves the metadata.  Then send an email to linux-btrfs.org explaining what you did and the errors you got and point to where somebody can download the image.  If Yan can use it he will pull it down.  Thanks,

Josef

Comment 7 Michael Hagmann 2010-07-07 21:15:09 UTC
Ok

thanks mike

Comment 8 Michael Hagmann 2010-07-13 20:29:44 UTC
hi Josef

sorry but the btrfs-image crashed

[root@enif data1]# btrfs-image /dev/mapper/vg_local-lv_virt Bugzilla612036-2.img
Segmentation fault
[root@enif data1]# 


I could only do a dd, does that also help?

thanks Mike

Comment 9 Michael Hagmann 2010-08-29 16:34:10 UTC
Just for Info

this bug was triggerd because of the disk where not 100% compatible with the 3ware controller 9690SA ( have to fix the linkspeed to 1.5 GB ) and there is a bug in the virtio driver for fs over 2 TB

https://bugzilla.redhat.com/show_bug.cgi?id=605757


whith the errata and the linkspeed fix, the problem is gone.

Michael

*** This bug has been marked as a duplicate of bug 605757 ***