Re: 'make check' failure with coreutils-8.2
Chris Clayton wrote: I've just built coreutils-8.2 and make check fails and asks for a report to this mail address. configure is run thusly: ./configure --prefix=/usr --disable-acl --disable-rpath --disable-nls --disable-xattr ... Thanks for the report. What type of system are you using? (uname -a) That failure means that your system's kernel lacks inotify support and that test failed to account for the possibility. Can you confirm that the patch below solves the problem? FAIL: tail-2/wait (exit: 1) === tail (GNU coreutils) 8.2 ... tail: cannot open `not_here' for reading: No such file or directory tail: inotify cannot be used, reverting to polling: Function not implemented tail: no files remaining + test 1 = 124 + test '!' -r unreadable + timeout 1 tail -s0.1 -f unreadable tail: cannot open `unreadable' for reading: Permission denied tail: inotify cannot be used, reverting to polling: Function not implemented tail: no files remaining + test 1 = 124 + timeout 1 tail -s0.1 -f here + test 124 = 124 + timeout 1 tail -s0.1 -F here + test 124 = 124 + test '!' -r unreadable + timeout 1 tail -s0.1 -F unreadable tail: cannot open `unreadable' for reading: Permission denied tail: inotify cannot be used, reverting to polling: Function not implemented + test 124 = 124 + timeout 1 tail -s0.1 -F not_here tail: cannot open `not_here' for reading: No such file or directory tail: inotify cannot be used, reverting to polling: Function not implemented + test 124 = 124 + test -s tail.err + fail=1 ... From 75db742ff9b0242db693c7e98c6adee265872cd0 Mon Sep 17 00:00:00 2001 From: Jim Meyering meyer...@redhat.com Date: Sat, 12 Dec 2009 14:44:46 +0100 Subject: [PATCH] tests: tail-without-inotify: avoid spurious test failure * tests/tail-2/wait: Account for the possibility that the kernel lacks inotify support. Reported by Chris Clayton. --- tests/tail-2/wait |3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/tests/tail-2/wait b/tests/tail-2/wait index 9392ade..bdc115a 100755 --- a/tests/tail-2/wait +++ b/tests/tail-2/wait @@ -54,7 +54,8 @@ for inotify in ---disable-inotify ''; do timeout 1 tail -s0.1 -F $inotify not_here test $? = 124 || fail=1 - + grep -v 'tail: inotify cannot be used, reverting to polling:' tail.err x + mv x tail.err test -s tail.err fail=1 :tail.err -- 1.6.6.rc1.319.g9b57d
Possible tail bug
Just noticed this having upgraded to Fedora 12. I use tail -F to follow a file that vanishes and is recreated once a minute. It works for some time but eventually tail quits and goes back to the command prompt. Just thought I'd let you know in case you changed something. tail: `info.log' has become inaccessible: No such file or directory tail: `info.log' has appeared; following end of new file ver-good budgetschooltours.ie ver-good 202.offersforme.net ver-good vistagy.com ver-good estatements.rbsnbonline.com ver-good antiqbook.com ver-bad ns207154.ovh.net ver-good ventech-eng.com tail: `info.log' has become inaccessible: No such file or directory tail: `info.log' has appeared; following end of new file ver-good wunderman.rjs0.com ver-good m.e.harryanddavid.com ver-good wunderman.rjs0.com ver-good cinci.rr.com ver-good tee.gr tail: `info.log' has been replaced; following end of new file ver-good mchsi.com ver-good joann-mail.com ver-good marplesteam.com ver-good tee.gr ver-good localnet.com tail: `info.log' has become inaccessible: No such file or directory [r...@listdata tmp]#
Re: Possible tail bug
Marc Perkel wrote: Just noticed this having upgraded to Fedora 12. I use tail -F to follow a file that vanishes and is recreated once a minute. It works for some time but eventually tail quits and goes back to the command prompt. Just thought I'd let you know in case you changed something. tail: `info.log' has become inaccessible: No such file or directory tail: `info.log' has appeared; following end of new file ver-good budgetschooltours.ie ver-good 202.offersforme.net ver-good vistagy.com ver-good estatements.rbsnbonline.com ver-good antiqbook.com ver-bad ns207154.ovh.net ver-good ventech-eng.com tail: `info.log' has become inaccessible: No such file or directory tail: `info.log' has appeared; following end of new file ver-good wunderman.rjs0.com ver-good m.e.harryanddavid.com ver-good wunderman.rjs0.com ver-good cinci.rr.com ver-good tee.gr tail: `info.log' has been replaced; following end of new file ver-good mchsi.com ver-good joann-mail.com ver-good marplesteam.com ver-good tee.gr ver-good localnet.com tail: `info.log' has become inaccessible: No such file or directory [r...@listdata tmp]# Thanks for the report. Can you reproduce that even with the very latest? coreutils-8.2 was released yesterday: ftp://ftp.gnu.org/gnu/coreutils/coreutils-8.2.tar.xz ftp://ftp.gnu.org/gnu/coreutils/coreutils-8.2.tar.gz and 8.1 includes a related fix: tail -F (inotify-enabled) now consistently tails a file that has been replaced via renaming. That operation provokes either of two sequences of inotify events. The less common sequence is now handled as well. [The bug came with the implementation change in coreutils-7.5] F12 is based on coreutils-7.6.
Fwd: 'make check' failure with coreutils-8.2
Added bug-coreutils@gnu.org, which I missed on the reply below... -- Forwarded message -- From: Chris Clayton chris2...@googlemail.com Date: 2009/12/12 Subject: Re: 'make check' failure with coreutils-8.2 To: Jim Meyering j...@meyering.net 2009/12/12 Jim Meyering j...@meyering.net: Chris Clayton wrote: I've just built coreutils-8.2 and make check fails and asks for a report to this mail address. configure is run thusly: ./configure --prefix=/usr --disable-acl --disable-rpath --disable-nls --disable-xattr ... Thanks for the report. What type of system are you using? (uname -a) It's a linux system running 2.6.32. Somehow, the kernel was set up to build without inotify support for userland. My 2.6.31 kernel has inotify support, so I've no idea how it got switched off. That failure means that your system's kernel lacks inotify support and that test failed to account for the possibility. Can you confirm that the patch below solves the problem? Yes, the patch allows the coreutils build to complete with the inotify-less kernel. I've rebuilt and installed my kernel too :-) Thanks Jim. Chris -- No, Sir; there is nothing which has yet been contrived by man, by which so much happiness is produced as by a good tavern or inn - Doctor Samuel Johnson -- No, Sir; there is nothing which has yet been contrived by man, by which so much happiness is produced as by a good tavern or inn - Doctor Samuel Johnson
tail + inotify over nfs
I've just noticed that `tail -f` will not work over NFS as changes on the remote system will not go through the local VFS and so will not be noticed by inotify. Testing with `tail -s.1 ---disable-inotify` does get the changes noticed, albeit after a delay of 3s on my setup. So what to do? I suppose we could statfs(filename) and then use polling for CIFS CODA FUSECTL? LUSTRE NFS SMB. I wonder should lib/mountlist.c::read_file_system_list() be updated as it only considers NFS, SMB and CIFS as remote? Also I wonder why GFS is not in the fs.h list? cheers, Pádraig.
