Skip to contentSkip to conversation

fstests: btrfs/271: specify "-m raid1" to avoid false alerts

PATCH In review Qu Wenruo 4 participants 5 messages Apr 17, 2025 – Apr 21, 2025 Btrfs file system

lore
Share & tools
lore

Qu Wenruo

author
#0 lore raw

PATCH 1/1 fstests: btrfs/271: specify "-m raid1" to avoid false alerts

[FALSE FAILURE]
Test case btrfs/271 will failure like the following, if the MKFS_OPTIONS has specified a metadata profile (either SINGLE or DUP):

btrfs/271 1s ... - output mismatch (see /home/adam/xfstests/results//btrfs/271.out.bad)
--- tests/btrfs/271.out 2022-11-07 09:59:11.256666666 +1030
+++ /home/adam/xfstests/results//btrfs/271.out.bad 2025-04-17 19:49:00.129443427 +0930
@@ -1,523 +1,9 @@
QA output created by 271
Allow global fail_make_request feature
Step 1: writing with one failing mirror:
-wrote 8192/8192 bytes at offset 0
-XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+fsync: Input/output error
Step 2: verify that the data reads back fine:
...
(Run 'diff -u /home/adam/xfstests/tests/btrfs/271.out /home/adam/xfstests/results//btrfs/271.out.bad' to see the entire diff)
Ran: btrfs/271
Failures: btrfs/271
Failed 1 of 1 tests

[CAUSE]
The test case relies on mkfs.btrfs to use RAID1 as default metadata profile if multiple devices are provided.

This is no longer true if the run has specified certain profile in MKFS_OPTIONS.

If "-m dup" or "-m single" is specified, the fs will flip read-only as either profile can handle any missing or failed device super block writeback.

[FIX]
Just specify both metadata (RAID1) and data (RAID1, already in the test case) to avoid the false failure.

---
 tests/btrfs/271 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tests/btrfs/271 +1 -1
index 2fc38e9c..7d5424f8 100755
@@ -18,7 +18,7 @@ _require_scratch_dev_pool 2
 _scratch_dev_pool_get 2
 
 _check_minimal_fs_size $(( 1024 * 1024 * 1024 ))
-_scratch_pool_mkfs "-d raid1 -b 1G" >> $seqres.full 2>&1
+_scratch_pool_mkfs "-m raid1 -d raid1 -b 1G" >> $seqres.full 2>&1
 
 _scratch_mount
 
Signature · 1 line
2.47.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 3 lines · Qu Wenruo

[FALSE FAILURE]
Test case btrfs/271 will failure like the following, if the MKFS_OPTIONS

will failure -> will fail

Quoted 30 lines

has specified a metadata profile (either SINGLE or DUP):

btrfs/271 1s ... - output mismatch (see /home/adam/xfstests/results//btrfs/271.out.bad)
--- tests/btrfs/271.out 2022-11-07 09:59:11.256666666 +1030
+++ /home/adam/xfstests/results//btrfs/271.out.bad 2025-04-17 19:49:00.129443427 +0930
@@ -1,523 +1,9 @@
QA output created by 271
Allow global fail_make_request feature
Step 1: writing with one failing mirror:
-wrote 8192/8192 bytes at offset 0
-XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+fsync: Input/output error
Step 2: verify that the data reads back fine:
...
(Run 'diff -u /home/adam/xfstests/tests/btrfs/271.out /home/adam/xfstests/results//btrfs/271.out.bad' to see the entire diff)
Ran: btrfs/271
Failures: btrfs/271
Failed 1 of 1 tests

[CAUSE]
The test case relies on mkfs.btrfs to use RAID1 as default metadata profile if multiple devices are provided.

This is no longer true if the run has specified certain profile in MKFS_OPTIONS.

If "-m dup" or "-m single" is specified, the fs will flip read-only as either profile can handle any missing or failed device super block writeback.

Quoted 6 lines

[FIX]
Just specify both metadata (RAID1) and data (RAID1, already in the test case) to avoid the false failure.

Looks good, thanks.

Quoted 21 lines
---
 tests/btrfs/271 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tests/btrfs/271 +1 -2
index 2fc38e9c..7d5424f8 100755
@@ -18,7 +18,7 @@ _require_scratch_dev_pool 2
 _scratch_dev_pool_get 2
 
 _check_minimal_fs_size $(( 1024 * 1024 * 1024 ))
-_scratch_pool_mkfs "-d raid1 -b 1G" >> $seqres.full 2>&1
+_scratch_pool_mkfs "-m raid1 -d raid1 -b 1G" >> $seqres.full 2>&1
 
 _scratch_mount
 
--

2.47.1

Reviewed-by
#2 lore raw
Quoted 6 lines · Filipe Manana
Quoted 3 lines · Qu Wenruo

[FALSE FAILURE]
Test case btrfs/271 will failure like the following, if the MKFS_OPTIONS

will failure -> will fail

Fixed. Applied. (for-next)

Thanks, Anand