Re: Btrfs broken in massive transfar

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




On 2018年03月13日 05:57, MASAKI haruka wrote:
> I'm trying to clone 18TiB data between btrfs,
> but it will crash anyway.
> 
> This probrem is occured even how to clone (btrfs send/receive, rsync or cp.)
> I experienced same probrem in Linux 3.9 and Linux 3.10.

Did you really mean *3*.9 and *3*.10?

That's too old for btrfs usage IIRC.

It would be *4*.9 or *4*.10 for a relative new kernel for btrfs.

Would you please try some latest mainline kernel again?

> 
> What happen:
> 
> 1. Failed to write because I/O error (read only filesystem)
> 2. writing to the btrfs succeeds and fails randomly.
> 3. The btrfs unable to unmount (resource is busy.) Unable to umount even forcely, so cannot halt.
> 
> Example:
> ---
> mkfile o7784-11-0
> rename o7784-11-0 -> .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo/kVxX8RdGhryQiEMOm4II2qMw
> utimes .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo
> truncate .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo/kVxX8RdGhryQiEMOm4II2qMw size=1073698824
> chown .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo/kVxX8RdGhryQiEMOm4II2qMw - uid=1000, gid=1000
> chmod .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo/kVxX8RdGhryQiEMOm4II2qMw - mode=0600
> utimes .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo/kVxX8RdGhryQiEMOm4II2qMw
> mkfile o7785-12-0
> rename o7785-12-0 -> .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo/lSABmfoArm9pAtade-gHmS6X
> utimes .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo
> truncate .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo/lSABmfoArm9pAtade-gHmS6X size=864067592
> ERROR: truncate .filesystem/HDD/.XFV_pp/,fQO40jotqhUZ0/5JSSubx1Ph5xYNOcXhIAoIK3/XDGOWpbx,5zYWEi0L5LHdWBo/lSABmfoArm9pAtade-gHmS6X failed: Input/output error
> btrfs send 180310235348  0.09s user 11.98s system 16% cpu 1:14.42 total
> ---

In that case, we need kernel message to investigate.
(And of course, please use at least 4.x kernel)

Thanks,
Qu

> 
> Tries:
> 1.
> Connect between host A (btrfs, 4disks) and B with socat (TCP).
> Host B write to iSCSI disk (btrfs, single).
> clone with btrfs send/receive. Linux 4.15.
> -> Crashed at transfarred 1.78TB
> 
> 2.
> Delete snapshot and retry.
> Connect between host A and B with SSH and socat (UNIX).
> Host B write to iSCSI disk (btrfs, single).
> clone with btrfs send/receive. Linux 4.15.
> -> Crashed at transfarred 90GB
> 
> 3.
> Recreate btrfs.
> Host A write to iSCSI disk.
> clone with btrfs send/receive. Linux 4.15.
> -> Crashed at transfarred 260GB
> 
> 4.
> Recreate btrfs.
> Original disk attach to other computer (having more resource.)
> clone with btrfs send/receive. Linux 4.15.
> -> Crashed at transfarred 120GB
> 
> 5.
> Recreate btrfs.
> Clone with rsync. Linux 4.15.
> -> Crashed at transfarred 100GB
> 
> 6.
> Recreate btrfs.
> Try with Linux 4.14, btrfs send/receive.
> -> Crashed at transfarred 3.98TB
> 
> 7.
> Recreate btrfs.
> Connect between host and NAS (iSCSI) with GbE cable directly.
> Mounted with options relatime, spase_cache, compress=lzo.
> clone with btrfs send/receive. Linux 4.14.
> -> Crashed at transfarred 2.13TB
> 

Attachment: signature.asc
Description: OpenPGP digital signature


[Index of Archives]     [Linux Filesystem Development]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux