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-5069 Lustre <p>on MDS2 hit the LBUG</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>LustreError: 2958:0:(mdd_dir.c:3954:mdd_migrate()) Skipped 15 previous similar messages LustreError: 2830:0:(lu_object.h:852:lu_object_attr()) ASSERTION( ((o)-&gt;lo_header-&gt;loh_attr &amp; LOHA_EXISTS) != 0 ) failed: LustreError: 2830:0:(lu_object.h:852:lu_object_attr()) LBUG Pid: 2830, comm: mdt00_003 Call Trace: Message from [&lt;ffffffffa0399895&gt;] libcfs_debug_dumpstack+0x55/0x80 [libcfs] syslogd@client-1 [&lt;ffffffffa0399e97&gt;] lbug_with_loc+0x47/0xb0 [libcfs] 8 at May 15 14:1 [&lt;ffffffffa0f3ae97&gt;] mdd_is_subdir+0x277/0x280 [mdd] 4:19 ... kern [&lt;ffffffffa0e0f2ef&gt;] mdt_rename_sanity+0xff/0x4a0 [mdt] el:LustreError: [&lt;ffffffffa0e1321c&gt;] mdt_reint_rename_internal+0xdc/0x1a80 [mdt] 2830:0:(lu_objec [&lt;ffffffffa06e46f8&gt;] ? ldlm_lock_enqueue+0x1c8/0x930 [ptlrpc] t.h:852:lu_objec [&lt;ffffffffa0703edb&gt;] ? ldlm_cli_enqueue_local+0x28b/0x5e0 [ptlrpc] t_attr()) ASSERT [&lt;ffffffffa0e14e04&gt;] mdt_reint_rename_or_migrate+0x244/0x660 [mdt] ION( ((o)-&gt;lo_he [&lt;ffffffffa0702bc0&gt;] ? ldlm_blocking_ast+0x0/0x180 [ptlrpc] ader-&gt;loh_attr &amp; [&lt;ffffffffa0704230&gt;] ? ldlm_completion_ast+0x0/0x930 [ptlrpc] LOHA_EXISTS) != [&lt;ffffffffa0e15250&gt;] mdt_reint_rename+0x10/0x20 [mdt] 0 ) failed: [&lt;ffffffffa0e0d881&gt;] mdt_reint_rec+0x41/0xe0 [mdt] [&lt;ffffffffa0df2e93&gt;] mdt_reint_internal+0x4c3/0x7c0 [mdt] [&lt;ffffffffa0df371b&gt;] mdt_reint+0x6b/0x120 [mdt] Message from [&lt;ffffffffa078fe5c&gt;] tgt_request_handle+0x23c/0xac0 [ptlrpc] syslogd@client- [&lt;ffffffffa073faea&gt;] ptlrpc_main+0xd1a/0x1980 [ptlrpc] 18 at May 15 14: [&lt;ffffffffa073edd0&gt;] ? ptlrpc_main+0x0/0x1980 [ptlrpc] 14:19 ... ker [&lt;ffffffff8109ab56&gt;] kthread+0x96/0xa0 [&lt;ffffffff8100c20a&gt;] child_rip+0xa/0x20 nel:LustreError: [&lt;ffffffff8109aac0&gt;] ? kthread+0x0/0xa0 [&lt;ffffffff8100c200&gt;] ? child_rip+0x0/0x20 2830:0:(lu_objeLustreError: dumping log to /tmp/lustre-log.1400188459.2830 ct.h:852:lu_object_attr()) LBUG </pre> </div></div> 2MDS with 4MDT<br/> 8OST<br/> 1 client<br/> build # <a href="http://build.whamcloud.com/job/lustre-reviews/23893/">http://build.whamcloud.com/job/lustre-reviews/23893/</a> LU-5069

Hit LBUG in DNE racer test: (lu_object.h:852:lu_object_attr()) ASSERTION( ((o)->lo_header->loh_attr & LOHA_EXISTS) != 0 ) failed

Bug Critical Resolved Fixed Di Wang Sarah Liu Thu, 15 May 2014 21:18:27 +0000 Tue, 3 Jun 2014 22:38:49 +0000 Tue, 3 Jun 2014 22:38:49 +0000 Lustre 2.6.0 Lustre 2.6.0 0 5 <p>It seems caused by this </p> <p> <a href="https://jira.whamcloud.com/browse/LU-4725" title="wrong lock ordering in rename leads to deadlocks" class="issue-link" data-issue-key="LU-4725"><del>LU-4725</del></a> mdt: child-parent lock ordering in rename</p> <p> change rename so that it always has parent-child lock ordering,<br/> otherwise it may deadlock with other operations.</p> <p> Signed-off-by: Vitaly Fertman &lt;vitaly_fertman@xyratex.com&gt;<br/> Signed-off-by: Hongchao Zhang &lt;hongchao.zhang@intel.com&gt;<br/> Change-Id: If676da82ca50a20a4bb3aadef0f81c9c5ed3cbcb<br/> Xyratex-bug-id: MRP-1700<br/> Reviewed-on: <a href="http://review.whamcloud.com/9538" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/9538</a><br/> Tested-by: Jenkins<br/> Tested-by: Maloo &lt;hpdd-maloo@intel.com&gt;<br/> Reviewed-by: wangdi &lt;di.wang@intel.com&gt;<br/> Reviewed-by: Oleg Drokin &lt;oleg.drokin@intel.com&gt;</p> <p>Hmm, do mdt_sanity_check without ldlm lock protection seems a bit risky. at least it needs to check whether object exist before mdo_is_subdir</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>static int mdt_rename_sanity(struct mdt_thread_info *info, const struct lu_fid *dir_fid, const struct lu_fid *fid) { ............... &lt;-------------------- check whether the object(dot) exists here. rc = mdo_is_subdir(info-&gt;mti_env, mdt_object_child(dst), fid, &amp;dst_fid); mdt_object_put(info-&gt;mti_env, dst); } </pre> </div></div> <p>I will cook a patch.</p> <p><a href="http://review.whamcloud.com/10340" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/10340</a></p> <p>Landed for 2.6</p> Development Rank 1|hzwmlr: Rank (Obsolete) 13997 Severity