Lustre discuss - Jan 2007 - [Fwd: Re: 1.4.8, shared root clusters and flock...]

Hi Oleg,

There are 45 tests in the fcntl section of the POSIX test suite.  I 
think around 10 of them require the the filesystem is mounted with 
flock, so I don''t know about "comprehensive", but
there''s a start.

- Richard


>-------- Original Message --------
>Subject:	Re: [Lustre-discuss] 1.4.8, shared root clusters and flock...
>Date:	Sat, 6 Jan 2007 02:11:56 +0200
>From:	Oleg Drokin <green@clusterfs.com>
>To:	David Golden <dgolden@cp.dias.ie>
>CC:	lustre-discuss@clusterfs.com
>References:	<200701051802.02805.dgolden@cp.dias.ie>
>
>
>
>Hello!
>
>On Fri, Jan 05, 2007 at 06:02:02PM +0000, David Golden wrote:
>
>>Just wondering is anyone else hitting this:
>>1.4.8 (correctly I guess!) started rejecting locks unless the 
>>"flock" mount option was used, as per changelog note for bug 
>>#10743.  I guess in the
>>past it faked success but the file wasn''t really locked?
>
>File was locked, but only on a that node, other nodes won''t see it.
>
>>=> So, what''s the latest on file lock support in general in
lustre?
>We are trying to fix when we learn about problems with it.
>(see bug 11415 if you plan to use fcntl locks with anything above 
>2.6.5 kernel)
>
>>Actually using the "flock" mount option doesn''t seem
to be a
>>solution, as AFAIK it''s known incomplete in 1.4.x (?): Unless
flock
>>is supposed to be working fully in 1.4.8, in which case, well, 
>>maybe there is a bug to be found...
>
>Well, we found one already.
>I wonder if there are any comprehensive flock/fcntl locking testsuites
>we can try to use?
>
>Bye,
>    Oleg
>
>_______________________________________________
>Lustre-discuss mailing list
>Lustre-discuss@clusterfs.com
>https://mail.clusterfs.com/mailman/listinfo/lustre-discuss

-- 



===================================================
Richard Hedges
Scalable I/O Project
Development Environment Group - Livermore Computing
Lawrence Livermore National Laboratory
7000 East Avenue, MS L-557
Livermore, CA	94551

v:	(925) 423-2699
f:	(925) 423-6961
E:	richard-hedges@llnl.gov

On Saturday 06 January 2007 00:34, Richard Hedges wrote:
> Hi Oleg,
>
> There are 45 tests in the fcntl section of the POSIX test suite.  I
> think around 10 of them require the the filesystem is mounted with
> flock, so I don''t know about "comprehensive", but
there''s a start.
I haven''t used or even looked at the open group''s test suites,
but I''m
guessing they only test intranode locking? (internode probably outside 
scope?)

A while back colleague forwarded me a tiny (definitely not comprehensive, but 
if the test fails, fs locking certainly can''t be used internode...)
perl
stress test for internode parallel locking, basically just do ~10K 
locks+writes per node on N nodes at once to a single file, and inspect if the 
file comes out gibberish or neat at the end. Lustre fs mounted without flock 
used to fail it (unsurprisingly) but I was going to give it another go 
when/if I get our cluster''s fs mounted with flock.

Hi,
I have found that when I mount via nfs a subdirectory of a lustre  
file system I get the message posted below (plus it doesnt work).

Example-

/disk/data

/disk is the lustre file system and /disk/data is the directory I  
want to export. I can export /disk just fine but I really only want  
to export /disk/data for some security reasons.

Message found in /var/log/messages--

LustreError: 4539:0:(ldlm_lock.c:82:ldlm_it2str()) Unknown intent  
-2145573515
LustreError: 4539:0:(dcache.c:290:ll_frob_intent()) ASSERTION(it- 
 >it_magic == INTENT_MAGIC) failed:bad intent magic: 802db80
LustreError: 4539:0:(dcache.c:290:ll_frob_intent()) LBUG
Lustre: 4539:0:(linux-debug.c:156:libcfs_debug_dumpstack()) showing  
stack for process 4539
rpc.mountd    R  running task       0  4539      1          4612   
4535 (NOTLB)
000001000117a030 ffffffff801d037b 0000000045a25af1 000000000b6bd680
        000000d272cf7110 0000000000000246 0000000000000256  
0000010000013780
        0000000000000000 00000000000000d2
Call Trace:<ffffffff801d037b>{inode_has_perm+89} <ffffffff801cf5b4> 
{avc_has_perm+70}
        <ffffffff801d2575>{selinux_file_permission+298}  
<ffffffffa0149a5d>{:sunrpc:cache_write+186}
        <ffffffff80177424>{vfs_write+207}
<ffffffff8017750c>{sys_write
+69}
        <ffffffff8011026a>{system_call+126}
LustreError: dumping log to /tmp/lustre-log-crew01.iges.org. 
1168268017.4539
Lustre: 4539:0:(linux-debug.c:96:libcfs_run_upcall()) Invoked LNET  
upcall /usr/lib/lustre/lnet_upcall LBUG,/testsuite/tmp/lbuild-boulder/ 
lbuild-v1_4_7_3-2.6-rhel4-x86_64/lbuild/BUILD/lustre-1.4.7.3/lustre/ 
llite/dcache.c,ll_frob_intent,290