Re: Balance loops: what we know so far

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

 




On 2020/5/5 上午2:54, Andrea Gelmini wrote:
> Il giorno mar 28 apr 2020 alle ore 12:00 Qu Wenruo
> <quwenruo.btrfs@xxxxxxx> ha scritto:
> 
>>
>> Would you please take an image dump of the fs when runaway balance happened?
>>
>> Both metadata and data block group loop would greatly help.
> 
> Hi Qu, and thanks a lot for your work.
> 
> I have a VirtualBox machine (in save-state) with the problem running.
> 
> So, I can send it to you, if you are comfortable with VirtulBox, or we
> can work together on my machine.

I mean, btrfs-image dump of the umounted fs.
(btrfs-image can compress the metadata, and won't include data, thus it
can be way smaller than the image)

> 
> Short story:
> a) system running is live CD ubuntu 20.04 with kernel 5.4.0-21
> b) partition of 10 giga (not /) mounted;
> c) shrinked to 4 giga;
> d) then removed files and left just 59 mega of data;
> e) started balance -m (not -d), then it loops forever (or, at least, I
> left it running for hours, guess it should be enough).

At this stage, the image should be pretty small.
You can try restart the system and boot into liveCD and dump the image.

It won't matter that much even you can't reproduce the loop after
restart, the point is to find out what's preventing the metadata balance
to continue. So the image-dump should be good enough.

Thanks,
Qu

> 
> If I reset it, at restart no more loop.
> 
> Thanks again,
> Gelma
> 

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