Skip to contentSkip to conversation

btrfs: fix resending received snapshot with parent

PATCH In review Robin Ruede 4 participants 5 messages Sep 30, 2015 – Jan 2, 2022 Btrfs file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 btrfs: fix resending received snapshot with parent

This fixes a regression introduced by 37b8d27d between v4.1 and v4.2.

When a snapshot is received, its received_uuid is set to the original uuid of the subvolume. When that snapshot is then resent to a third filesystem, it's received_uuid is set to the second uuid instead of the original one. The same was true for the parent_uuid. This behaviour was partially changed in 37b8d27d, but in that patch only the parent_uuid was taken from the real original,
not the uuid itself, causing the search for the parent to fail in the case below.

This happens for example when trying to send a series of linked
snapshots (e.g. created by snapper) from the backup file system back to the original one.

The following commands reproduce the issue in v4.2.1 (no error in 4.1.6)

    # setup three test file systems
    for i in 1 2 3; do
	    truncate -s 50M fs$i
	    mkfs.btrfs fs$i
	    mkdir $i
	    mount fs$i $i
    done
    echo "content" > 1/testfile
    btrfs su snapshot -r 1/ 1/snap1
    echo "changed content" > 1/testfile
    btrfs su snapshot -r 1/ 1/snap2
    # works fine:
    btrfs send 1/snap1 | btrfs receive 2/
    btrfs send -p 1/snap1 1/snap2 | btrfs receive 2/

# ERROR: could not find parent subvolume btrfs send 2/snap1 | btrfs receive 3/ btrfs send -p 2/snap1 2/snap2 | btrfs receive 3/

---
 fs/btrfs/send.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
fs/btrfs/send.c +8 -2
index aa72bfd..890933b 100644
@@ -2351,8 +2351,14 @@ static int send_subvol_begin(struct send_ctx *sctx)
 	}
 
 	TLV_PUT_STRING(sctx, BTRFS_SEND_A_PATH, name, namelen);
-	TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
-			sctx->send_root->root_item.uuid);
+
+	if (!btrfs_is_empty_uuid(sctx->send_root->root_item.received_uuid))
+		TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
+			    sctx->send_root->root_item.received_uuid);
+	else
+		TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
+			    sctx->send_root->root_item.uuid);
+
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_CTRANSID,
 		    le64_to_cpu(sctx->send_root->root_item.ctransid));
 	if (parent_root) {
Signature · 1 line
2.6.0
Reviewed-by on v1 1/1
#1 lore raw
Quoted 33 lines · Robin Ruede

This fixes a regression introduced by 37b8d27d between v4.1 and v4.2.

When a snapshot is received, its received_uuid is set to the original uuid of the subvolume. When that snapshot is then resent to a third filesystem, it's received_uuid is set to the second uuid instead of the original one. The same was true for the parent_uuid. This behaviour was partially changed in 37b8d27d, but in that patch only the parent_uuid was taken from the real original,
not the uuid itself, causing the search for the parent to fail in the case below.

This happens for example when trying to send a series of linked
snapshots (e.g. created by snapper) from the backup file system back to the original one.

The following commands reproduce the issue in v4.2.1 (no error in 4.1.6)

    # setup three test file systems
    for i in 1 2 3; do
            truncate -s 50M fs$i
            mkfs.btrfs fs$i
            mkdir $i
            mount fs$i $i
    done
    echo "content" > 1/testfile
    btrfs su snapshot -r 1/ 1/snap1
    echo "changed content" > 1/testfile
    btrfs su snapshot -r 1/ 1/snap2
    # works fine:
    btrfs send 1/snap1 | btrfs receive 2/
    btrfs send -p 1/snap1 1/snap2 | btrfs receive 2/
Quoted 5 lines

# ERROR: could not find parent subvolume btrfs send 2/snap1 | btrfs receive 3/ btrfs send -p 2/snap1 2/snap2 | btrfs receive 3/

Thanks for fixing this.
I've added this to my integration branch [1] and will send soon a pull request to Chris for 4.4, including this fix plus a few others for send/receive, after some more testing.

I've also made an xfstest for it [1, 2]

[1] http://git.kernel.org/cgit/linux/kernel/git/fdmanana/linux.git/log/?h=integration-4.3
[2] https://patchwork.kernel.org/patch/7363681/
[3] https://patchwork.kernel.org/patch/7363661/

Quoted 32 lines
---
 fs/btrfs/send.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
fs/btrfs/send.c +8 -3
index aa72bfd..890933b 100644
@@ -2351,8 +2351,14 @@ static int send_subvol_begin(struct send_ctx *sctx)
        }
 
        TLV_PUT_STRING(sctx, BTRFS_SEND_A_PATH, name, namelen);
-       TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
-                       sctx->send_root->root_item.uuid);
+
+       if (!btrfs_is_empty_uuid(sctx->send_root->root_item.received_uuid))
+               TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
+                           sctx->send_root->root_item.received_uuid);
+       else
+               TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
+                           sctx->send_root->root_item.uuid);
+
        TLV_PUT_U64(sctx, BTRFS_SEND_A_CTRANSID,
                    le64_to_cpu(sctx->send_root->root_item.ctransid));
        if (parent_root) {
--

2.6.0

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Signature · 5 lines
Filipe David Manana,

"Reasonable men adapt themselves to the world.
 Unreasonable men adapt the world to themselves.
 That's why all progress depends on unreasonable men."
Tested-by
#2 lore raw
Quoted 7 lines · Filipe Manana
Quoted 6 lines · Robin Ruede

This fixes a regression introduced by 37b8d27d between v4.1 and v4.2.

When a snapshot is received, its received_uuid is set to the original uuid of the subvolume. When that snapshot is then resent to a third filesystem, it's received_uuid is set to the second uuid instead of the original one. The same was true for the parent_uuid. ...

Quoted 8 lines

Thanks for fixing this.
I've added this to my integration branch [1] and will send soon a pull request to Chris for 4.4, including this fix plus a few others for send/receive, after some more testing.

I've also made an xfstest for it [1, 2]

Another thanks for this fix. It fixes things here. I am runing 4.2.3 with the 4.3 btrfs tree pulled on top of it along with this fix. Incremental sends are now working again.

This fixes a regression, can we please get into 4.3?

Thanks
Ed Tomlinson

#3 lore raw
Quoted 21 lines · Ed Tomlinson
Quoted 10 lines · Filipe Manana
Quoted 7 lines · Robin Ruede

This fixes a regression introduced by 37b8d27d between v4.1 and v4.2.

When a snapshot is received, its received_uuid is set to the original uuid of the subvolume. When that snapshot is then resent to a third filesystem, it's received_uuid is set to the second uuid instead of the original one. The same was true for the parent_uuid. ...

Quoted 8 lines

Thanks for fixing this.
I've added this to my integration branch [1] and will send soon a pull request to Chris for 4.4, including this fix plus a few others for send/receive, after some more testing.

I've also made an xfstest for it [1, 2]

Quoted 7 lines

Another thanks for this fix. It fixes things here. I am runing 4.2.3 with the 4.3 btrfs tree pulled on top of it along with this fix. Incremental sends
are now working again.

This fixes a regression, can we please get into 4.3?

I've tagged it for stable backports in my 4.4 integration branch [1]. Thanks.

[1] http://git.kernel.org/cgit/linux/kernel/git/fdmanana/linux.git/log/?h=integration-4.4

thanks

Quoted 4 lines

Thanks
Ed Tomlinson

Signature · 5 lines
Filipe David Manana,

"Reasonable men adapt themselves to the world.
 Unreasonable men adapt the world to themselves.
 That's why all progress depends on unreasonable men."
#4 lore raw
In reply to Robin Ruede

trešd., 2015. g. 30. sept., plkst. 22:25 — lietotājs Robin Ruede (<[email protected]>) rakstīja:

Quoted 67 lines

This fixes a regression introduced by 37b8d27d between v4.1 and v4.2.

When a snapshot is received, its received_uuid is set to the original uuid of the subvolume. When that snapshot is then resent to a third filesystem, it's received_uuid is set to the second uuid instead of the original one. The same was true for the parent_uuid. This behaviour was partially changed in 37b8d27d, but in that patch only the parent_uuid was taken from the real original,
not the uuid itself, causing the search for the parent to fail in the case below.

This happens for example when trying to send a series of linked
snapshots (e.g. created by snapper) from the backup file system back to the original one.

The following commands reproduce the issue in v4.2.1 (no error in 4.1.6)

    # setup three test file systems
    for i in 1 2 3; do
            truncate -s 50M fs$i
            mkfs.btrfs fs$i
            mkdir $i
            mount fs$i $i
    done
    echo "content" > 1/testfile
    btrfs su snapshot -r 1/ 1/snap1
    echo "changed content" > 1/testfile
    btrfs su snapshot -r 1/ 1/snap2
    # works fine:
    btrfs send 1/snap1 | btrfs receive 2/
    btrfs send -p 1/snap1 1/snap2 | btrfs receive 2/

# ERROR: could not find parent subvolume btrfs send 2/snap1 | btrfs receive 3/ btrfs send -p 2/snap1 2/snap2 | btrfs receive 3/

---
 fs/btrfs/send.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
fs/btrfs/send.c +8 -3
index aa72bfd..890933b 100644
@@ -2351,8 +2351,14 @@ static int send_subvol_begin(struct send_ctx *sctx)
        }
 
        TLV_PUT_STRING(sctx, BTRFS_SEND_A_PATH, name, namelen);
-       TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
-                       sctx->send_root->root_item.uuid);
+
+       if (!btrfs_is_empty_uuid(sctx->send_root->root_item.received_uuid))
+               TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
+                           sctx->send_root->root_item.received_uuid);
+       else
+               TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
+                           sctx->send_root->root_item.uuid);
+
        TLV_PUT_U64(sctx, BTRFS_SEND_A_CTRANSID,
                    le64_to_cpu(sctx->send_root->root_item.ctransid));
        if (parent_root) {
--

2.6.0

Quoted 1 line

--

Hi,

This breaks use case when snapshot is sent from one system to another where it's set to read-write, modified, set back to read-only to send to 3rd system.
This will cause received_uuid to be present but it won't match original first snapshot anymore so 2nd system's UUID should be used instead of first.
Only very recently (Oct 2021) "btrfs property set" was changed to clear received_uuid when switching ro -> rw thus it means there could be a lot of rw filesystems with it set. And it can cause strange send failure, see my email titled "btrfs send picks wrong subvolume UUID". Also I'm not only one who encountered such issue [1]. Workaround is to clear received_uuid, but that might not be obvious or not even possible for filesystems mounted as read-only (or corrupted ones that can't be remounted rw).

[1] "Wrong received UUIDs after backup restore and onward" -
https://www.spinics.net/lists/linux-btrfs/msg65142.html

Best regards,
Dāvis