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-14640 Lustre <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> Lustre: DEBUG MARKER: == sanity-benchmark test fsx: fsx ==================================================================== 22:56:45 (1619287005) LustreError: 8995:0:(osd_io.c:1563:osd_write_commit()) ASSERTION( !PageDirty(lnb[i].lnb_page) ) failed: LustreError: 8995:0:(osd_io.c:1563:osd_write_commit()) LBUG Pid: 8995, comm: ll_ost_io00_009 4.18.0 #36 SMP Thu Mar 25 14:56:29 MSK 2021 Call Trace: [&lt;0&gt;] libcfs_call_trace+0x76/0xa0 [libcfs] [&lt;0&gt;] lbug_with_loc+0x3e/0x80 [libcfs] [&lt;0&gt;] osd_write_commit+0x33a/0x9a0 [osd_ldiskfs] [&lt;0&gt;] ofd_commitrw+0xbb2/0x2d90 [ofd] [&lt;0&gt;] tgt_brw_write+0xeb7/0x22f0 [ptlrpc] [&lt;0&gt;] tgt_request_handle+0xbe0/0x1970 [ptlrpc] [&lt;0&gt;] ptlrpc_main+0x134f/0x30e0 [ptlrpc] </pre> </div></div> <p>bisected to:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java">cb037f305c 2020-12-14 | LU-14160 fallocate: Add punch mode to fallocate [Arshad Hussain]</pre> </div></div> LU-14640

ASSERTION( !PageDirty(lnb[i].lnb_page) in osd_write_commit()

Bug Minor Resolved Fixed Arshad Hussain Alex Zhuravlev Sun, 25 Apr 2021 05:46:11 +0000 Sat, 22 Jan 2022 08:22:02 +0000 Tue, 31 Aug 2021 11:16:52 +0000 Upstream Lustre 2.15.0 0 5 <p>Hi Arshad, could you please take a look. </p> <p>I suspect this is being caused by <tt>fallocate(PUNCH_HOLE)</tt> on non-page boundaries, which is leaving the partially-zeroed page in the buffer cache. With osd-ldiskfs we don't leave any pages in the buffer cache, so that we don't need to invalidate them later when doing large direct read/write operations. These buffers (potentially at both the start and end of the punch) should be treated the same way as a regular truncate where it zeroes the partial page, submits it for write, then invalidates the buffer afterward. </p> <p>Hi Andreas, I am looking into this...</p> <p>Hi Andreas,</p> <p>&gt; ... These buffers (potentially at both the start and end of the punch) should be treated the same way as a regular truncate where it zeroes the partial page, submits it for write, then invalidates the buffer afterward. </p> <p>Please correct me. It looks like the fallocate(PUNCH) is currently doing partial page flush.</p> <p> </p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> osd_trans_stop() ----&gt;osd_process_truncates() --------&gt;osd_execute_punch()              {                      osd_partial_page_flush(start); ---&gt; calls (filemap_fdatawrite_range)                      osd_partial_page_flush(end); ---&gt; calls (filemap_fdatawrite_range)              } </pre> </div></div> <p>Unfortunalely, on my local setup I could not get sanity-benchmark(fsx) test to reproduce the assertion yet even running in a loop.</p> <p>I am forcing a partial punch for verification  via script. This is is also passing.</p> <p> </p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>#!/bin/bash fname=/mnt/lustre/A yes 'a' | dd of=$fname bs=4096 count=100 sync; sleep 1 for i in {100..1}; do offset=$(($i * 4096)) length=4090 echo "Running: fallocate -p -o $offset -l $length $fname" fallocate -p -o $offset -l $length $fname done </pre> </div></div> <p> </p> <p> </p> <p>Hi Alex,</p> <p>I could not get sanity-benchmark(fsx) test to reproduce the assertion on my local setup. Could you help me with your setup details ? I am trying to reproduct this on the current master (a2b5290 <a href="https://jira.whamcloud.com/browse/LU-14613" title="sanity/60f fails on a local setup" class="issue-link" data-issue-key="LU-14613"><del>LU-14613</del></a> tests: fix sanity/60f on a local setup)</p> <p>nothing really special - /tmp in ram (tmpfs), then REFORMAT=y ONLY=fsx ONLY_REPEAT=1000 sh sanity-benchmark.sh</p> <p>&gt;nothing really special - /tmp in ram (tmpfs), then REFORMAT=y ONLY=fsx ONLY_REPEAT=1000 sh sanity-benchmark.sh</p> <p>I will try out with increased loop count. Thanks!</p> <p>Arshad Hussain (arshad.hussain@aeoncomputing.com) uploaded a new patch: <a href="https://review.whamcloud.com/43462" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/43462</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-14640" title="ASSERTION( !PageDirty(lnb[i].lnb_page) in osd_write_commit()" class="issue-link" data-issue-key="LU-14640"><del>LU-14640</del></a> osd: non aligned flush<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 48a492c9d87163dc8c55a4ad1042bbab9a07448c</p> <p>"Oleg Drokin &lt;green@whamcloud.com&gt;" merged in patch <a href="https://review.whamcloud.com/43462/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/43462/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-14640" title="ASSERTION( !PageDirty(lnb[i].lnb_page) in osd_write_commit()" class="issue-link" data-issue-key="LU-14640"><del>LU-14640</del></a> osd: ASSERTION(!PageDirty(lnb<span class="error">&#91;i&#93;</span>.lnb_page)<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 76c5e8ed9560fe232bcc0c2ee0069dbdb8411565</p> <p>Landed for 2.15</p> Related LU-14402 Development Rank 1|i01t1r: Rank (Obsolete) 9223372036854775807 Severity