Skip to contentSkip to conversation

ERROR: invalid size for attribute, expected = 4, got = 8 on subvolume send

DISCUSSION Nemcev Aleksey 1 participant 2 messages Nov 3, 2022 Btrfs file system

lore
Share & tools
lore
#0 lore raw

Hello.

 I'm using Btrfs with zstd compression (subvolume property compression=zstd) on the PC and on the backup drive and running as a daemon the bees (on-demand deduplication tool) on the PC.

 I'm using btrfs send | btrfs receive to send incremental snapshots of PC subvolumes and receive them to the backup drive. Sometimes receive fails with ERROR: invalid size for attribute, expected = 4, got = 8 or ERROR: invalid size for attribute, expected = 8, got = 4

 Why does this error happen?
 How can I debug this error? At least how to get the problem file name?
 Is that attribute an internal Btrfs-specific thing or something generic like xattr attribute or file attribute (lsattr/chattr)? Is it possible to resolve this error without removing the subvolume and snapshot on the backup drive and resending the complete subvolume again?

 Thank you.

#1 lore raw

03.11.2022 12:32, Nemcev Aleksey пишет:

Quoted 7 lines

Hello.

 I'm using Btrfs with zstd compression (subvolume property compression=zstd) on the PC and on the backup drive and running as a daemon the bees (on-demand deduplication tool) on the PC.

 I'm using btrfs send | btrfs receive to send incremental snapshots of PC subvolumes and receive them to the backup drive. Sometimes receive fails with ERROR: invalid size for attribute, expected = 4, got = 8 or ERROR: invalid size for attribute, expected = 8, got = 4

Quoted 6 lines

 Why does this error happen?
 How can I debug this error? At least how to get the problem file name?
 Is that attribute an internal Btrfs-specific thing or something generic like xattr attribute or file attribute (lsattr/chattr)? Is it possible to resolve this error without removing the subvolume and snapshot on the backup drive and resending the complete subvolume again?

 Thank you.

UPD:

It seems that the --compressed-data option on btrfs-send was the issue, when I removed it, and now btrfs-receive works fine on the same PC and backup drive. This is not an issue for me, as I'm using bash pipe to connect btrfs-send and btrfs-receive (but this wastes CPU time for decompression and recompression data during send/receive). But this can be an issue for anyone with a slow channel between send and receive sides.