Dear All, I try to use our cluster though samba share. Everything work fine, but I think, we should have -o flock at lustre mount time. Great, it''s work. But when I want to save a file on the share, I get this on the logs: Jan 31 10:45:24 opteron-ren-11 kernel: LustreError: 24836:0:(file.c:2309:ll_file_flock()) unknown fcntl lock type: 32 Jan 31 10:45:24 opteron-ren-11 kernel: LustreError: 24836:0:(file.c:2310:ll_file_flock()) LBUG Jan 31 10:45:24 opteron-ren-11 kernel: Lustre: 24836:0:(linux-debug.c:171:libcfs_debug_dumpstack()) can''t show stack: kernel doesn''t export show_task Jan 31 10:45:24 opteron-ren-11 kernel: Jan 31 10:45:24 opteron-ren-11 kernel: Call Trace: Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff8857ad0a>] :libcfs:lbug_with_loc+0x7a/0xc0 Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff887c2a65>] :lustre:ll_file_flock+0x295/0x550 Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff8025fa48>] do_page_fault+0x476/0x7b5 Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff8868fba0>] :ptlrpc:ldlm_flock_completion_ast+0x0/0x690 Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff802ac2ff>] audit_syscall_entry+0x141/0x174 Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff802d39d6>] sys_flock+0x117/0x150 Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff8025729c>] tracesys+0xdc/0xe1 Jan 31 10:45:24 opteron-ren-11 kernel: Jan 31 10:45:24 opteron-ren-11 kernel: LustreError: dumping log to /tmp/lustre-log.1201772724.24836 I attached the lustre-log. What could cause this problem? The reason, I wanted to use flock is there are some softwares, which cannot save files by various reasons on the cluster share (for example MS O2K3, Project 2K7). But as I can see, right now doesn''t matter, it is a notepad, or a Project 2K7. Cluster''s version is 1.6.4.1 and 1.6.0.1 and I tested this with 1.6.4.1 and 1.6.2. What can I do? Thank you, tamas -------------- next part -------------- A non-text attachment was scrubbed... Name: lustre-log.zip Type: application/zip Size: 125760 bytes Desc: not available Url : http://lists.lustre.org/pipermail/lustre-discuss/attachments/20080131/6609dad1/attachment-0002.zip
On Thu, 2008-01-31 at 11:10 +0100, Papp Tamas wrote:> Dear All, > > I try to use our cluster though samba share. Everything work fine, but > I think, we should have -o flock at lustre mount time.Is this a single samba server exporting Lustre? Do you really need lock coherency between Lustre clients? If the answers are yes and no, you''d be better to use -o localflock. b. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part Url : http://lists.lustre.org/pipermail/lustre-discuss/attachments/20080131/5c64a514/attachment-0002.bin
Tamas, I?ve had issues with software that used POSIX file locking calls with Lustre 1.6 ... the same thing worked on Lustre 1.4, but I suspect it just failed transparently and didn?t actually involve any recording locking. Never saw it panic anything, though, but I know that 1.6 doesn?t support locking (yet). Klaus On 1/31/08 2:10 AM, "Papp Tamas" <tompos at martos.bme.hu>did etch on stone tablets:> Dear All, > > I try to use our cluster though samba share. Everything work fine, but > I think, we should have -o flock at lustre mount time. > > Great, it''s work. But when I want to save a file on the share, I get > this on the logs: > > Jan 31 10:45:24 opteron-ren-11 kernel: LustreError: > 24836:0:(file.c:2309:ll_file_flock()) unknown fcntl lock type: 32 > Jan 31 10:45:24 opteron-ren-11 kernel: LustreError: > 24836:0:(file.c:2310:ll_file_flock()) LBUG > Jan 31 10:45:24 opteron-ren-11 kernel: Lustre: > 24836:0:(linux-debug.c:171:libcfs_debug_dumpstack()) can''t show stack: kernel > doesn''t export show_task > Jan 31 10:45:24 opteron-ren-11 kernel: > Jan 31 10:45:24 opteron-ren-11 kernel: Call Trace: > Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff8857ad0a>] > :libcfs:lbug_with_loc+0x7a/0xc0 > Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff887c2a65>] > :lustre:ll_file_flock+0x295/0x550 > Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff8025fa48>] > do_page_fault+0x476/0x7b5 > Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff8868fba0>] > :ptlrpc:ldlm_flock_completion_ast+0x0/0x690 > Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff802ac2ff>] > audit_syscall_entry+0x141/0x174 > Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff802d39d6>] > sys_flock+0x117/0x150 > Jan 31 10:45:24 opteron-ren-11 kernel: [<ffffffff8025729c>] > tracesys+0xdc/0xe1 > Jan 31 10:45:24 opteron-ren-11 kernel: > Jan 31 10:45:24 opteron-ren-11 kernel: LustreError: dumping log to > /tmp/lustre-log.1201772724.24836 > > > I attached the lustre-log. > > What could cause this problem? The reason, I wanted to use flock is > there are some softwares, which cannot save files by various reasons > on the cluster share (for example MS O2K3, Project 2K7). > > But as I can see, right now doesn''t matter, it is a notepad, or a > Project 2K7. > > > Cluster''s version is 1.6.4.1 and 1.6.0.1 and I tested this with > 1.6.4.1 and 1.6.2. > > > What can I do? > > > Thank you, > > tamas >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.lustre.org/pipermail/lustre-discuss/attachments/20080131/43cdab9b/attachment-0002.html