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-1563 Lustre <p>Our customer hit the following LBUG when probably the quota is enabled. <br/> It looks like this is very similar to <a href="https://jira.whamcloud.com/browse/LU-1224" title="LBUG ASSERTION(atomic_read(&amp;lqs-&gt;lqs_refcount) &gt; 0) failed" class="issue-link" data-issue-key="LU-1224"><del>LU-1224</del></a>, but it's not fixed even yet.</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>Jun 13 17:03:57 nos021i kernel: LustreError: 8625:0:(lustre_quota.h:326:lqs_putref()) ASSERTION(atomic_read(&amp;lqs-&gt;lqs_refcount) &gt; 0) failed Jun 13 17:03:57 nos021i kernel: LustreError: 8625:0:(lustre_quota.h:326:lqs_putref()) LBUG Jun 13 17:03:57 nos021i kernel: Pid: 8625, comm: ll_ost_io_21 Jun 13 17:03:57 nos021i kernel: Jun 13 17:03:57 nos021i kernel: Call Trace: Jun 13 17:03:57 nos021i kernel: [&lt;ffffffff886fe6a1&gt;] libcfs_debug_dumpstack+0x51/0x60 [libcfs] Jun 13 17:03:57 nos021i kernel: [&lt;ffffffff886febda&gt;] lbug_with_loc+0x7a/0xd0 [libcfs] Jun 13 17:03:57 nos021i kernel: [&lt;ffffffff88706fc0&gt;] tracefile_init+0x0/0x110 [libcfs] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff889572ef&gt;] quota_pending_commit+0x41f/0x5b0 [lquota] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff800a2dff&gt;] autoremove_wake_function+0x0/0x2e Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88b4690b&gt;] fsfilt_ldiskfs_commit_wait+0xab/0xd0 [fsfilt_ldiskfs] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88b8786f&gt;] filter_commitrw_write+0x253f/0x2dd0 [obdfilter] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88b21e1a&gt;] ost_checksum_bulk+0x2aa/0x5a0 [ost] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88b28d09&gt;] ost_brw_write+0x1c99/0x2480 [ost] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88862ac8&gt;] ptlrpc_send_reply+0x5e8/0x600 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff8882d8b0&gt;] target_committed_to_req+0x40/0x120 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff8008e7f9&gt;] default_wake_function+0x0/0xe Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff888670a8&gt;] lustre_msg_check_version_v2+0x8/0x20 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88b2c09e&gt;] ost_handle+0x2bae/0x55b0 [ost] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff887b8d00&gt;] class_handle2object+0xe0/0x170 [obdclass] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff8882119a&gt;] lock_res_and_lock+0xba/0xd0 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88826168&gt;] __ldlm_handle2lock+0x2f8/0x360 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff888766d9&gt;] ptlrpc_server_handle_request+0x989/0xe00 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88876e35&gt;] ptlrpc_wait_event+0x2e5/0x310 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff8008cc1e&gt;] __wake_up_common+0x3e/0x68 Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88877dc6&gt;] ptlrpc_main+0xf66/0x1120 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff8005dfb1&gt;] child_rip+0xa/0x11 Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff88876e60&gt;] ptlrpc_main+0x0/0x1120 [ptlrpc] Jun 13 17:03:58 nos021i kernel: [&lt;ffffffff8005dfa7&gt;] child_rip+0x0/0x11 </pre> </div></div> lustre-1.8.7-wc1, RHEL5(Server), RHEL6(Client) LU-1563

lustre_quota.h:326:lqs_putref() LBUG

Bug Minor Resolved Fixed Niu Yawei Shuichi Ihara Mon, 25 Jun 2012 11:24:53 +0000 Fri, 22 Feb 2013 11:18:50 +0000 Mon, 27 Aug 2012 01:29:16 +0000 Lustre 1.8.7 Lustre 1.8.9 0 6 <p>OSS's log files are uploaded on /uploads/<a href="https://jira.whamcloud.com/browse/LU-1563" title="lustre_quota.h:326:lqs_putref() LBUG" class="issue-link" data-issue-key="LU-1563"><del>LU-1563</del></a>. We hit LBUG on these three OSSs.</p> <p>FYI, I suspect it's same with BZ24188, however the patch on BZ24188 is for 2.x which has a total new implementation of cfs_hash, so we can't make a direct backport.</p> <p>Niu</p> <p>Could you please look into this one?</p> <p>Thanks</p> <p>Peter</p> <p>This part of code is racy, quota_pending_commit() could find a lqs (not being held in quota_check_common()) then put it twice.<br/> I posted a patch for b1_8: <a href="http://review.whamcloud.com/#change,3187" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/#change,3187</a> </p> <p>Niu</p> <p>Will this same change be needed for master?</p> <p>Peter</p> <p>Hi, Peter</p> <p>No, the code in master is different from b1_8, it has similar checking in the quota_pending_commit() like this patch. Thanks.</p> <p>ok. Then marking as resolved as this fix has landed to b1_8. We can reopen if the fix transpires not to have fixed the reported problem</p> Development Rank 1|hzvgt3: Rank (Obsolete) 6372 Severity