[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: tail aborts while following by name if using inotify
From: |
Jim Meyering |
Subject: |
Re: tail aborts while following by name if using inotify |
Date: |
Wed, 30 Dec 2009 11:26:38 +0100 |
Giuseppe Scrivano wrote:
>> However, while looking at it, I discovered another problem.
>> When tail-F'd files may be renamed, tail may fail to track
>> the target of a rename.
>
> good catch and very useful test case! I am going to like the test
> driven development we we had today :-)
>
> This patch should fix the problem, other tests remain green.
Looks good. Thanks!
I've adjusted the log and pushed that, followed by the new
test and a NEWS update:
>From 2ead2365e01f8e103ddce6f34558eee3648de5b4 Mon Sep 17 00:00:00 2001
From: Giuseppe Scrivano <address@hidden>
Date: Wed, 30 Dec 2009 00:20:24 +0100
Subject: [PATCH 1/3] tail -F: don't stop following the target of a rename
This fixes a bug whereby tail -F would fail to track changes
to a file that was a target of a rename, and when the source of
the rename was another tailed file.
* src/tail.c (tail_forever_inotify): Ensure the wd is not already
present in the hash table before trying to add it. When a new watch
descriptor is added to the `wd_to_name' hash table, check that it is
not already present. If it is present then remove the previous element.
---
src/tail.c | 15 ++++++++++++++-
1 files changed, 14 insertions(+), 1 deletions(-)
diff --git a/src/tail.c b/src/tail.c
index 3d5e221..28a0e26 100644
--- a/src/tail.c
+++ b/src/tail.c
@@ -1486,11 +1486,24 @@ tail_forever_inotify (int wd, struct File_spec *f,
size_t n_files,
/* Remove `fspec' and re-add it using `new_fd' as its key. */
hash_delete (wd_to_name, fspec);
fspec->wd = new_wd;
+
+ /* If the file was moved then inotify will use the source file wd for
+ the destination file. Make sure the key is not present in the
+ table. */
+ struct File_spec *prev = hash_delete (wd_to_name, fspec);
+ if (prev && prev != fspec)
+ {
+ if (follow_mode == Follow_name)
+ recheck (prev, false);
+ prev->wd = -1;
+ close_fd (prev->fd, pretty_name (prev));
+ }
+
if (hash_insert (wd_to_name, fspec) == NULL)
xalloc_die ();
if (follow_mode == Follow_name)
- recheck (&(f[j]), false);
+ recheck (fspec, false);
}
else
{
--
1.6.6.rc3.271.g3d40f
>From fe062d563a09481162f5c11b154d047331708b7f Mon Sep 17 00:00:00 2001
From: Jim Meyering <address@hidden>
Date: Tue, 29 Dec 2009 16:37:04 +0100
Subject: [PATCH 2/3] tail: test for a bug in inotify-enabled tail -F
tail -F a b would stop tracking additions to b after "mv a b".
* tests/tail-2/F-vs-rename: New file.
* tests/Makefile.am (TESTS): Add it.
---
tests/Makefile.am | 1 +
tests/tail-2/F-vs-rename | 78 ++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 79 insertions(+), 0 deletions(-)
create mode 100755 tests/tail-2/F-vs-rename
diff --git a/tests/Makefile.am b/tests/Makefile.am
index 6ef7ad8..33b9df4 100644
--- a/tests/Makefile.am
+++ b/tests/Makefile.am
@@ -82,6 +82,7 @@ TESTS = \
cp/link-heap \
tail-2/inotify-hash-abuse \
tail-2/inotify-hash-abuse2 \
+ tail-2/F-vs-rename \
tail-2/inotify-rotate \
chmod/no-x \
chgrp/basic \
diff --git a/tests/tail-2/F-vs-rename b/tests/tail-2/F-vs-rename
new file mode 100755
index 0000000..75d3c69
--- /dev/null
+++ b/tests/tail-2/F-vs-rename
@@ -0,0 +1,78 @@
+#!/bin/sh
+# demonstrate that tail -F works when renaming the tailed files
+# Before coreutils-8.3, tail -F a b would stop tracking additions to b
+# after "mv a b".
+
+# Copyright (C) 2009 Free Software Foundation, Inc.
+
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+if test "$VERBOSE" = yes; then
+ set -x
+ tail --version
+fi
+
+. $srcdir/test-lib.sh
+
+touch a b || framework_failure
+
+debug='---disable-inotify -s .01'
+debug=
+tail $debug -F a b > out 2>&1 & pid=$!
+
+# Wait until tail has started...
+echo x > a
+until grep '^x$' out >/dev/null 2>&1; do :; done
+
+mv a b || fail=1
+
+# Wait for the diagnostic:
+# tail: `a' has become inaccessible: No such file or directory
+until grep inaccessible out >/dev/null 2>&1; do :; done
+
+echo x > a
+# Wait up to 4s for this to appear in the output:
+# "tail: `...' has appeared; following end of new file"
+found=false
+for i in $(seq 20); do
+ grep 'has appeared;' out > /dev/null && { found=true; break; }
+ sleep .2
+done
+$found || { echo "$0: a: unexpected delay?"; cat out; fail=1; }
+
+echo y >> b
+# Wait up to 4s for "y" to appear in the output:
+found=false
+for i in $(seq 20); do
+ case $(tr '\n' @ < out) in
+ *'@@==> b <address@hidden@') found=true; break 2;;
+ esac
+ sleep .2
+done
+$found || { echo "$0: b: unexpected delay?"; cat out; fail=1; }
+
+echo z >> a
+# Wait up to 4s for "z" to appear in the output:
+found=false
+for i in $(seq 20); do
+ case $(tr '\n' @ < out) in
+ *'@@==> a <address@hidden@') found=true; break 2;;
+ esac
+ sleep .2
+done
+$found || { echo "$0: b: unexpected delay?"; cat out; fail=1; }
+
+kill -HUP $pid
+
+Exit $fail
--
1.6.6.rc3.271.g3d40f
>From d3a4bc86cd989ad7262545ae7e49868a2d1b59d4 Mon Sep 17 00:00:00 2001
From: Jim Meyering <address@hidden>
Date: Wed, 30 Dec 2009 11:16:23 +0100
Subject: [PATCH 3/3] doc: mention two tail -F bug fixes in NEWS
* NEWS (Bug fixes): Two tail -F fixes.
---
NEWS | 9 +++++++++
1 files changed, 9 insertions(+), 0 deletions(-)
diff --git a/NEWS b/NEWS
index 6128e3f..dd2ad6d 100644
--- a/NEWS
+++ b/NEWS
@@ -18,6 +18,15 @@ GNU coreutils NEWS -*-
outline -*-
files that was not done from the local system would go unnoticed.
[bug introduced in coreutils-7.5]
+ tail -F (inotify-enabled) would abort when a tailed file is repeated
+ renamed-aside and then recreated.
+ [bug introduced in coreutils-7.5]
+
+ tail -F (inotify-enabled) could fail to follow renamed files.
+ E.g., given a "tail -F a b" process, running "mv a b" would
+ make tail stop tracking additions to "b".
+ [bug introduced in coreutils-7.5]
+
touch -a once again guarantees that a file's change time is
adjusted, working around a bug in current Linux kernels.
[bug introduced in coreutils-8.1]
--
1.6.6.rc3.271.g3d40f