Windows reports drive not initialized, wants to create MBR/GPT about btrfs HOT 13 CLOSED

I think the driver's loaded - unless you've changed your partition type, it
wouldn't appear at all otherwise. Could you send me a debug log please?
There's instructions in the README.

On 2 Sep 2016 8:20 a.m., "Proberts" [email protected] wrote:

I'm on Windows 8.1 X64 build 9600. Test Mode is enabled. When I plug in my
BTRFS drive, Windows tells me it's not initialized and wants to put a
partition table on the drive. The drive mounts under Linux without issue.

I've tried both running loader.exe and installing with the inf and
rebooting. I don't see the a service named 'btrfs' in Computer Management
-> Services or Task Manager -> Services, however running loader.exe tells
me the service already exists and is already running.

Thanks for any help. Can't wait to start using btrfs on Windows!

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#29, or mute the thread
https://github.com/notifications/unsubscribe-auth/AI3sj8D2jlgYD1YlWEHN_OjK3ZzGR3fWks5ql8AygaJpZM4JzZvG
.

from btrfs.

Thanks for your reply. I found the problem - at least with the 8GB drive. The problem turned out to be between the monitor and keyboard. (The drive had been reformatted XFS and I didn't check it first.)

However, I was unable to access another drive I tried (which I did verify was btrfs - single volume, LZO enabled). Your driver seems to be picking it up and says it's mounted in the btrfs registry, I just can't find how to access it. I tried restarting Explorer and entering some drive letters (starting with G) into an explorer window. I've attached the log and some relevant screenshots.

5TB-btrfs.zip

from btrfs.

That's weird. I think it might be getting confused by your RAID0 drives for some reason - I'll have to experiment. The log says it's mounting successfully, but then no I/O is taking place - which is odd, because at the very least you'd expect Windows to look for $RECYCLE.BIN. Let me get back to you.

from btrfs.

Thanks for your reply. I'm using the RAID right now for work project, but if I finish before you get a chance to look at the issue, I'll try unplugging the RAID drives from the motherboard and report back with a log.

from btrfs.

So I'm afraid I can't get winbtrfs to recognize a btrfs volume now for some reason. I create the volume with defaults under Linux: mkfs.btrfs -L "TestWinBtrfs" /dev/sdX

When I attach the drive to Windows, Disk Manager says it's uninitialized (I've tried initializing in Windows, the reformatting on Linux with mkfs.btrfs) and the btrfs log loops this message:

0000000000000720:drv_file_system_control:file system control
0000000000000720:drv_file_system_control:IRP_MN_MOUNT_VOLUME
0000000000000720:mount_vol:mount_vol called
0000000000000720:mount_vol:partition length = 1cd93fc00
0000000000000720:read_superblock:got superblock 0!
0000000000000720:read_superblock:got superblock 1!
0000000000000720:read_superblock:crc32 was 08aed57a, expected ffffffff
0000000000000720:mount_vol:mount_vol done (status: c000014f)

Do I need to format with specific parameters? (This is a single volume external USB drive)

from btrfs.

I'm not sure, I've not tested partition-less mode with USB yet. I'll have to get back to you. That partition length is suspect though - it's saying it's 7,743,994,880 bytes, which isn't a round number. Is that really the case?

from btrfs.

Actually, I see what the issue is - mkfs.btrfs has only written one superblock, which is what the Linux driver does with things it thinks are SSDs. The Windows driver is then dying because it thinks the second superblock is corrupt, when really it should just be ignoring it. I'll get this fixed for the next version.

from btrfs.

@maharmstone Could this have anything to do with issue #19 given that device is an SSD device (mkfs.btrfs used 'single' for metadata). If so, is there anything you need from me to confirm the behaviour?

from btrfs.

I think it might, but let me check.

Can you try this please? Mount the volume on Linux, and so some sort of change, e.g. create a directory. Wait thirty seconds for the FS to be flushed, then do another change, then unmount. That'll update both superblocks, and I think it ought to work on Windows after that.

from btrfs.

Hi,

I just wanted to give some feedback with my case, that it's similar to what @Proberts is describing.
In my case, I have both a 2TB internal disk and a pendrive of 16GB.

In both cases the driver is working perfectly and I see the volumes mounted in explorer, but, anytime I get into the disk manager, windows says that the drive is uninitialized and I should add a GPT or MBR.

The funny thing is that I can use the drive without any problem, and even mount a specific subvol. No problems what's so ever. The only "problem" is that I cannot handle the driver letter assigned to the volumes.

from btrfs.

The problems with diskmgmt.msc seem to be due to Windows not liking the MBR/GPT partition type that Linux uses by default. Try changing it to 7 in fdisk.

I'll see if there's some way we can fake this automatically; if we can't, I'll add a note to the readme.

from btrfs.

With btrfs one can either format a single partition (then the disk is partitioned most likely using MBR/GPT), but there's also the possibility to format the whole disk as btrfs, without using any partition table. Linux supports that well and so does WinBtrfs. But diskmgmt then shows the alert, because there's no partition table present on the drive. In this situation, there are three options:

1. Live with it as is. Just don't accidentially click "initialize disk" in diskmgmt; in the worst case the btrfs formatting should be recoverable by filling the disk's first 65 kB with zeros (I involuntarily tested this already).
2. Clean the whole disk, partition it and create a new btrfs partition.
3. Move the filesystem on the disk, add a partition table and insert the filesystem.

from btrfs.

Closing old issues

from btrfs.