https://jira.whamcloud.com This file is an XML representation of an issue en-us 9.4.14 940014 05-12-2023 https://jira.whamcloud.com/browse/LU-4627 Lustre <p>While investigating <a href="https://jira.whamcloud.com/browse/LU-3732" title="osd_io.c:320:osd_do_bio()) ASSERTION( iobuf-&gt;dr_rw == 0 ) failed: page_idx 4, block_idx 4, i 0" class="issue-link" data-issue-key="LU-3732"><del>LU-3732</del></a> I've regularly bumped into what seems to be a deadlock on lli_trunc_sem, according to the trace below: </p> <p>Call Trace:<br/> <span class="error">&#91;&lt;ffffffff81511655&gt;&#93;</span> rwsem_down_failed_common+0x95/0x1d0<br/> <span class="error">&#91;&lt;ffffffff815117b3&gt;&#93;</span> rwsem_down_write_failed+0x23/0x30<br/> <span class="error">&#91;&lt;ffffffff81284163&gt;&#93;</span> call_rwsem_down_write_failed+0x13/0x20<br/> <span class="error">&#91;&lt;ffffffff81510cb2&gt;&#93;</span> ? down_write+0x32/0x40<br/> <span class="error">&#91;&lt;ffffffffa0e03fe1&gt;&#93;</span> ll_setattr_raw+0x191/0x10c0 <span class="error">&#91;lustre&#93;</span><br/> <span class="error">&#91;&lt;ffffffff810758c7&gt;&#93;</span> ? current_fs_time+0x27/0x30<br/> <span class="error">&#91;&lt;ffffffffa0e04f6d&gt;&#93;</span> ll_setattr+0x5d/0xf0 <span class="error">&#91;lustre&#93;</span><br/> <span class="error">&#91;&lt;ffffffff8119ec18&gt;&#93;</span> notify_change+0x168/0x340<br/> <span class="error">&#91;&lt;ffffffffa06bb3bb&gt;&#93;</span> ? libcfs_debug_vmsg2+0x50b/0xbb0 <span class="error">&#91;libcfs&#93;</span><br/> <span class="error">&#91;&lt;ffffffff811197ef&gt;&#93;</span> file_remove_suid+0x5f/0x90<br/> <span class="error">&#91;&lt;ffffffff8111c2c0&gt;&#93;</span> __generic_file_aio_write+0x220/0x490<br/> <span class="error">&#91;&lt;ffffffffa06bbaa1&gt;&#93;</span> ? libcfs_debug_msg+0x41/0x50 <span class="error">&#91;libcfs&#93;</span><br/> <span class="error">&#91;&lt;ffffffff8111c5b8&gt;&#93;</span> generic_file_aio_write+0x88/0x100<br/> <span class="error">&#91;&lt;ffffffffa0e4353b&gt;&#93;</span> vvp_io_write_start+0xdb/0x3d0 <span class="error">&#91;lustre&#93;</span><br/> <span class="error">&#91;&lt;ffffffffa08cbcaa&gt;&#93;</span> cl_io_start+0x6a/0x140 <span class="error">&#91;obdclass&#93;</span><br/> <span class="error">&#91;&lt;ffffffffa08cfe24&gt;&#93;</span> cl_io_loop+0xb4/0x1b0 <span class="error">&#91;obdclass&#93;</span><br/> <span class="error">&#91;&lt;ffffffffa0de2a36&gt;&#93;</span> ll_file_io_generic+0x2b6/0x710 <span class="error">&#91;lustre&#93;</span><br/> <span class="error">&#91;&lt;ffffffffa08bfdf9&gt;&#93;</span> ? cl_env_get+0x29/0x350 <span class="error">&#91;obdclass&#93;</span><br/> <span class="error">&#91;&lt;ffffffffa0de3702&gt;&#93;</span> ll_file_aio_write+0x142/0x2c0 <span class="error">&#91;lustre&#93;</span><br/> <span class="error">&#91;&lt;ffffffffa0de39ec&gt;&#93;</span> ll_file_write+0x16c/0x2a0 <span class="error">&#91;lustre&#93;</span><br/> <span class="error">&#91;&lt;ffffffff811814b8&gt;&#93;</span> vfs_write+0xb8/0x1a0<br/> <span class="error">&#91;&lt;ffffffff81181e72&gt;&#93;</span> sys_pwrite64+0x82/0xa0<br/> <span class="error">&#91;&lt;ffffffff8100b072&gt;&#93;</span> system_call_fastpath+0x16/0x1b</p> <p>It seems to me that the down_read() in ll_file_io_generic() should only be done if iot == CIT_READ since in the write case, we down_write() later.</p> LU-4627

Client deadlock on ll_setattr_raw

Bug Blocker Resolved Fixed Zhenyu Xu Henri Doreau Thu, 13 Feb 2014 13:04:36 +0000 Tue, 12 Jul 2016 17:39:26 +0000 Mon, 10 Mar 2014 23:45:53 +0000 Lustre 2.5.0 Lustre 2.6.0 Lustre 2.5.4 0 10 <p>Bobijam</p> <p>Could you please advise on this one?</p> <p>Thanks</p> <p>Peter</p> <p>What's the version of Lustre being used here?</p> <p>This is the current HEAD</p> <p>I think you are right about it. patch tracking at <a href="http://review.whamcloud.com/9267" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/9267</a></p> <p>Landed for 2.6</p> <p>Patch backport for b2_5 at <a href="http://review.whamcloud.com/#/c/12268" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/#/c/12268</a>. I'm hoping this will fix some performance issues with truncates I have seen.</p> <p>Oleg Drokin (oleg.drokin@intel.com) merged in patch <a href="http://review.whamcloud.com/12268/" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/12268/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-4627" title="Client deadlock on ll_setattr_raw" class="issue-link" data-issue-key="LU-4627"><del>LU-4627</del></a> llite: deed taking lli_trunc_sem during file write<br/> Project: fs/lustre-release<br/> Branch: b2_5<br/> Current Patch Set: <br/> Commit: 8a2fe616a959f18a928d4edf185467cbf905c355</p> Duplicate LU-4710 Related Development Rank 1|hzwf1j: Rank (Obsolete) 12657 Severity