generic/084 try to run 'tail' command, tail will use inotify.
There're some limit about the number of inotify. For example
fs.inotify.max_user_instances specifies an upper limit on
the number of inotify instances that can be created per real
user ID.
When I test on a machine with 154 cpu cores, this case run
failed, and hit many warning likes:
+tail: inotify cannot be used, reverting to polling: Too many open files
Because the fs.inotify.max_user_instances is 128, so if we
try to tail 154 files, it will be failed.
So use src/multi_open_unlink to instead of tail will avoid
this problem.
Signed-off-by: Zorro Lang <zlang@xxxxxxxxxx>
Reviewed-by: Eryu Guan <eguan@xxxxxxxxxx>
---
v1:
- use both fs.inotify's max_user_watches and max_user_instances
to decide the $nr_cpu number.
V2:
- remove max_user_watches, only use max_user_instances.
- add more commends to explain my change.
v3:
- follow Dave's suggestion, use multi_open_unlink to replace tail.
- test on old kernel, sure the bug can be reproduced after change
V4:
- remove -s 1000 from multi_open_unlink
- because Dave haven't ack+, so remove "Reviewed-by Dave"
tests/generic/084 | 15 ++++-----------
1 file changed, 4 insertions(+), 11 deletions(-)
diff --git a/tests/generic/084 b/tests/generic/084
index 3fec6c2..3324bc9 100755
--- a/tests/generic/084
+++ b/tests/generic/084
@@ -69,15 +69,8 @@ _scratch_mkfs >>$seqres.full 2>&1
_scratch_mount
# create, open & unlinked files so unlinked inode list is not empty
-for i in `seq 1 $nr_cpu`; do
- testfile=$SCRATCH_MNT/$seq.unlinked.$i
- touch $testfile
- tail -f $testfile &
- tail_pids="$tail_pids $!"
-done
-# give tail some time to open the file before the file is removed
-sleep 1
-rm -f $SCRATCH_MNT/$seq.unlinked.*
+src/multi_open_unlink -f $SCRATCH_MNT/$seq.unlinked -n $nr_cpu &
+open_pid=$!
# start link/unlink storm
src=$SCRATCH_MNT/$seq.target
@@ -96,8 +89,8 @@ done &
sleep 5
kill $! >/dev/null 2>&1
-kill $tail_pids $link_pids >/dev/null 2>&1
-wait $tail_pids $link_pids
+kill $open_pid $link_pids >/dev/null 2>&1
+wait $open_pid $link_pids
# all done, no oops/hang expected, _check_filesystems checks SCRATCH_DEV after test
status=0
--
1.9.3
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
