Recently my file server stops working (Centos 7.7). The btrfs vol "clouddata" unable to mount. The server hangs during mounting it in read-write mode. I thought space exhausted on it:

    [root@cloud auto]# btrfs fi show
    Label: 'clouddata'  uuid: 90224df7-9d85-48f7-8cb5-a6689fc4f0a5
    Total devices 1 FS bytes used 12.42TiB
    devid    1 size 14.45TiB used 12.84TiB path /dev/sda5

14.45 TB disk is hardware RAID10 on DELL R740xd. All drives are OK.

Mount options:

    /dev/sda5 /data01 btrfs compress=lzo,discard,clear_cache,subvol=data01   0 2

I can mount it in read-only. Other partitions using ext4 file system. Btrfs was chosen for "clouddata" because of snapshots . Server log-files:

    [  169.678015] BTRFS: error (device sda5) in maybe_insert_hole:4859: errno=-17 Object already exists
    [  169.678102] BTRFS info (device sda5): forced readonly
    [  169.678168] BTRFS warning (device sda5): Skipping commit of aborted transaction.
    [  169.678178] BTRFS: error (device sda5) in cleanup_transaction:1851: errno=-17 Object already exists

    [root@cloud cifs]# btrfs --version
    btrfs-progs v4.9.1

The command "btrfs check --repair /dev/sda5" failed with:

    Feb  2 22:58:09 cloud kernel: Out of memory: Kill process 2321 (btrfs) score 946 or sacrifice child
    Feb  2 22:58:09 cloud kernel: Killed process 2321 (btrfs), UID 0, total-vm:23160652kB, anon-rss:15259688kB, file-rss:176kB, shmem-rss:0kB

Is it free space problem? How to delete all snapshots on unmounted volume? There were no power outage.


Zero-log resolved this situation. However btrfs-wiki not recommend using it. I'm switching back to zfs on FreeBSD.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.