Has anyone else seen or had issues with local-btrfs. I have tested with every setting I can find. When trying to backup a VM or container that is using local-btrfs. It is not usable. I am talking 8-128kib/s I have adjusted vzdump I have tried no compression. You name it. I move the VM or CT to Ceph and no issues great performance on the backup.
I only had one VM I wanted to run in local-btrfs. PBS. Just to have it outside of the Ceph storage and make it easier to restore if Ceph needed to be rebuilt. But the performance is so bad using anything in local-btrfs. I have tested with ext4, zfs, and LVM with no issues. I have tested with new installs they all perform the same.
Has anyone found a solution to this? I have 10GB ethernet between all boxes with. 10GB to Ceph. I have normal backup performance as long as I am not trying to use anything in local-btrfs. All drives are 10k SAS no hardware raid. 6 drives in local-btrfs and 14 in Ceph. I have even tested with VM stopped, suspended or snapshots. I have setup options to set the bwlimit to 8881 and different speeds. Which performs great for ext4, zfs, LVM or Ceph when doing any migration, clone, storage, restores or backups. I noticed even PBS is slow to backup anything in local-btrfs.
I have another setup with SSD drives and 40G network and I see the same issues. I have been testing and trying to work around this for a month doing different tests. I am ready to give up on btrfs and go to a different local file system. I just want to be able to do snapshots and fast enough backups at night.
Oh, and the other issue I would see with vzdump and local-btrfs is it would error timeout and never complete any backups no matter how slow or fast I adjust.
I only had one VM I wanted to run in local-btrfs. PBS. Just to have it outside of the Ceph storage and make it easier to restore if Ceph needed to be rebuilt. But the performance is so bad using anything in local-btrfs. I have tested with ext4, zfs, and LVM with no issues. I have tested with new installs they all perform the same.
Has anyone found a solution to this? I have 10GB ethernet between all boxes with. 10GB to Ceph. I have normal backup performance as long as I am not trying to use anything in local-btrfs. All drives are 10k SAS no hardware raid. 6 drives in local-btrfs and 14 in Ceph. I have even tested with VM stopped, suspended or snapshots. I have setup options to set the bwlimit to 8881 and different speeds. Which performs great for ext4, zfs, LVM or Ceph when doing any migration, clone, storage, restores or backups. I noticed even PBS is slow to backup anything in local-btrfs.
I have another setup with SSD drives and 40G network and I see the same issues. I have been testing and trying to work around this for a month doing different tests. I am ready to give up on btrfs and go to a different local file system. I just want to be able to do snapshots and fast enough backups at night.
Oh, and the other issue I would see with vzdump and local-btrfs is it would error timeout and never complete any backups no matter how slow or fast I adjust.
Last edited: