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-12222 Lustre <p>I happened to notice that when mounting an MGT and MDT co-located on the same node that we end up sending RPCs to ourselves:</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>00000100:00000040:11.0:1556165595.786175:0:30688:0:(niobuf.c:905:ptl_send_rpc()) @@@ send flg=0 req@ffff881f275f8000 x1631757890748560/t0(0) o502-&gt;MGC10.12.0.50@o2ib40@10.12.0.50@o2ib40:26/25 lens 272/8472 e 0 to 0 dl 1556165690 ref 3 fl Rpc:/0/ffffffff rc 0/-1 00000100:00000001:11.0:1556165595.786177:0:30688:0:(niobuf.c:54:ptl_send_buf()) Process entered 00000100:00000040:11.0:1556165595.786177:0:30688:0:(niobuf.c:57:ptl_send_buf()) peer_id 12345-10.12.0.50@o2ib40 00000400:00000010:11.0:1556165595.786178:0:30688:0:(lib-lnet.h:221:lnet_md_alloc()) slab-alloced 'md' of size 136 at ffff881f279d26e8. 00000100:00000200:11.0:1556165595.786178:0:30688:0:(niobuf.c:85:ptl_send_buf()) Sending 272 bytes to portal 26, xid 1631757890748560, offset 0 00000400:00000010:11.0:1556165595.786179:0:30688:0:(lib-lnet.h:472:lnet_msg_alloc()) alloc '(msg)': 440 at ffff881f279c3200 (tot 331800384). 00000400:00000200:11.0:1556165595.786181:0:30688:0:(lib-move.c:4637:LNetPut()) LNetPut -&gt; 12345-10.12.0.50@o2ib40 00000400:00000200:11.0:1556165595.786183:0:30688:0:(lib-move.c:2545:lnet_handle_send_case_locked()) Source ANY to MR: 10.12.0.50@o2ib40 local destination 00000400:00000200:11.0:1556165595.786184:0:30688:0:(lib-move.c:1619:lnet_get_best_ni()) compare ni 0@lo [c:0, d:10, s:0] with best_ni not seleced [c:-2147483648, d:-1, s:0] 00000400:00000200:11.0:1556165595.786185:0:30688:0:(lib-move.c:1662:lnet_get_best_ni()) selected best_ni 0@lo 00000400:00000200:11.0:1556165595.786186:0:30688:0:(lib-move.c:1619:lnet_get_best_ni()) compare ni 10.12.0.50@o2ib40 [c:2045, d:10, s:26] with best_ni 0@lo [c:0, d:10, s:0] 00000400:00000200:11.0:1556165595.786188:0:30688:0:(lib-move.c:1662:lnet_get_best_ni()) selected best_ni 10.12.0.50@o2ib40 00000400:00000200:11.0:1556165595.786188:0:30688:0:(lib-move.c:1367:lnet_select_peer_ni()) 10.12.0.50@o2ib40 ni_is_pref = 0 00000400:00000200:11.0:1556165595.786189:0:30688:0:(lib-move.c:1428:lnet_select_peer_ni()) sd_best_lpni = 10.12.0.50@o2ib40 00000400:00000200:11.0:1556165595.786192:0:30688:0:(lib-move.c:1837:lnet_handle_send()) TRACE: 10.12.0.50@o2ib40(10.12.0.50@o2ib40:&lt;?&gt;) -&gt; 10.12.0.50@o2ib40(10.12.0.50@o2ib40:10.12.0.50@o2ib40) : PUT try# 0</pre> </div></div> <p>In the above excerpt we can see that LNet has enough information to determine that we're sending this to ourselves (the nid of best_ni == the nid of best_lpni). Ideally we'd just use the loopback NI whenever possible.</p> LU-12222

LNet should select loopback NI when possible

Improvement Major Resolved Fixed Alexey Lyashkov Chris Horn HPv3 Thu, 25 Apr 2019 04:46:25 +0000 Wed, 9 Dec 2020 21:08:16 +0000 Wed, 20 May 2020 13:57:12 +0000 Lustre 2.12.0 Lustre 2.13.0 Lustre 2.14.0 Lustre 2.12.6 0 8 <p>This seems pretty important. It could cause problems with local communication on the MDS or OSS under load. </p> <p>Chris,</p> <p>Are you planning a patch for this?</p> <p>I was, but Amir tells me that he relies on this behavior for recovery. If he can provide some more details on what he means by that then I can work out a patch, otherwise I'm fine to let Amir work it out himself.</p> <p>OK, thanks!  Was just curious.</p> <p>Chris Horn (hornc@cray.com) uploaded a new patch: <a href="https://review.whamcloud.com/35778" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/35778</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Check if we're sending to ourselves<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: ee8e5be8951ed60d25b3aeae563a0f54cde59ea7</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/35778/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/35778/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Check if we're sending to ourselves<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: e4af756e1f428a9f7883bf883f66941defb1447f</p> <p>Landed for 2.14</p> <p>In fact this is invalid fix, with performance impact.<br/> based on hits provided by Chirs, I able to replicate this issue and it looks like a ptlrpc bug.</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>00000400:00000200:5.0:1577100140.229954:0:10627:0:(peer.c:1215:LNetPrimaryNID()) NID 0@lo primary NID 198.18.0.33@tcp rc 0 00000100:00000010:5.0:1577100140.229963:0:10627:0:(connection.c:55:ptlrpc_connection_get()) kmalloced '(conn)': 88 at ffff942f2df04518. 00000100:00000001:5.0:1577100140.229965:0:10627:0:(connection.c:80:ptlrpc_connection_get()) Process leaving 00000100:00000040:5.0:1577100140.229966:0:10627:0:(connection.c:84:ptlrpc_connection_get()) conn=ffff942f2df04518 refcount 2 to 198.18.0.33@tcp 00000100:00000040:5.0:1577100140.229967:0:10627:0:(client.c:107:ptlrpc_uuid_to_connection()) MGC198.18.0.33@tcp_0 -&gt; ffff942f2df04518 </pre> </div></div> <p>ptlrpc_uuid_to_connection - don't able to understand this is local NID and provide a replacement as needs.</p> <p>In fact this MR bug.<br/> ptlrpc_uuid_to_peer - correctly resolved 198.18.0.33@tcp as local address, and it replaced to 0@lo, but <br/> ptlrpc_connection_get have call a LNetPrimaryNID on top, and it call make this replacement away.</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=pjones" class="user-hover" rel="pjones">pjones</a> <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=green" class="user-hover" rel="green">green</a> Since my fix for this issue has a bad interaction with the fix for <a href="https://jira.whamcloud.com/browse/LU-12889" title="Do not assume peers are MR capable" class="issue-link" data-issue-key="LU-12889"><del>LU-12889</del></a> I think we should revert it. The <a href="https://jira.whamcloud.com/browse/LU-12889" title="Do not assume peers are MR capable" class="issue-link" data-issue-key="LU-12889"><del>LU-12889</del></a> bug is more serious, and per Shadow's last couple comments in this ticket, he thinks we can resolve this issue in a different way. What do you think?</p> <p>since <a href="https://jira.whamcloud.com/browse/LU-12889" title="Do not assume peers are MR capable" class="issue-link" data-issue-key="LU-12889"><del>LU-12889</del></a> is not yet landed, you can submit a series consisting of reversal and then <a href="https://jira.whamcloud.com/browse/LU-12889" title="Do not assume peers are MR capable" class="issue-link" data-issue-key="LU-12889"><del>LU-12889</del></a> on top of it (or a single patch that reverts just the needed bits if desirable) with necessary explanations and all.</p> <p>Okay, that sounds like a plan. Thanks.</p> <p>Chris Horn (hornc@cray.com) uploaded a new patch: <a href="https://review.whamcloud.com/37259" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/37259</a><br/> Subject: Revert "<a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Check if we're sending to ourselves"<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: ee971ecfbeb4a8192e4ab0e2762048a89a31e339</p> <blockquote> <p>In fact this MR bug.<br/> ptlrpc_uuid_to_peer - correctly resolved 198.18.0.33@tcp as local address, and it replaced to 0@lo, but<br/> ptlrpc_connection_get have call a LNetPrimaryNID on top, and it call make this replacement away.</p></blockquote> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=shadow" class="user-hover" rel="shadow">shadow</a> - thanks for the clear analysis. We definitely don't want to be sending messages destined for ourselves out over the network. I'm not sure what the semantics for <tt>LNetPrimaryNID()</tt> are. Should the <tt>0@lo</tt> NID be handled specifically in that function, or would it be better to skip that function in <tt>ptlrpc_connection_get()</tt> and keep <tt>0@lo</tt> as the target NID? </p> <p>Andreas, From my view, exceptions should be on lower level as possible - to avoid similar bugs in future, in case ptlrpc code will changed and someone forget why this exceptions was made.</p> <p>Shadow, I was thinking that <tt>LNetPrimaryNID()</tt> might be called to map <tt>0@lo</tt> to the primary external NID for other reasons. However, it looks like there is only one caller of this function, in <tt>ptlrpc_connection_get()</tt>, so changing the semantic to just return <tt>0@lo</tt> immediately is pretty safe. I don't mind to add a fast path in both <tt>LNetPrimaryNID()</tt> and <tt>ptlrpc_connection_get()</tt> that special-case this NID to avoid overhead (e.g. skipping <tt>lnet_net_lock_current()</tt> and maybe avoiding the need to look up the peer connection?).</p> <p>It probably also makes sense to <tt>#define LNET_NID_LO_0</tt> (or similar) like we do for <tt>LNET_NID_ANY</tt> so that we don't have to compose this every time. That could also be used in other places, like <tt>infra_ping_nid()</tt> rather than turning it into a string and comparing that much more efficiently in that loop:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> - <span class="code-keyword">if</span> (!strcmp(libcfs_nid2str(ping.ping_buf[i].nid), - <span class="code-quote">"0@lo"</span>)) + <span class="code-keyword">if</span> (ping.ping_buf[i].nid == LNET_NID_LO_0)) </pre> </div></div> <p>Any comments from LNet developers on a global <tt>LNET_NID_LO_0</tt> constant to indicate "myself"?</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/37259/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/37259/</a><br/> Subject: Revert "<a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Check if we're sending to ourselves"<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 275016fcae729f8467dc09aa15aa2b4bb2690875</p> <blockquote><p>Any comments from LNet developers on a global LNET_NID_LO_0 constant to indicate "myself"?</p></blockquote> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=adilger" class="user-hover" rel="adilger">adilger</a> Sounds like a good idea to me.</p> <p>Chris Horn (chris.horn@hpe.com) uploaded a new patch: <a href="https://review.whamcloud.com/38312" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38312</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Introduce constant for the lolnd NID<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 620d738ec490c1fc475f6e2312f2290800051705</p> <p>Chris Horn (chris.horn@hpe.com) uploaded a new patch: <a href="https://review.whamcloud.com/38313" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38313</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Primary NID of lolnd NID is the lolnd NID<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: c11c50152cffea9e05fc8124f7b988fae828f1e4</p> <p>Chris Horn (chris.horn@hpe.com) uploaded a new patch: <a href="https://review.whamcloud.com/38388" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38388</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> ptlrpc: Check if NID is local, not just lolnd NID<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: b17862a23870bc3be8f24ae9441dab316cfac611</p> <p>Without the fix, performing llmount.sh with -1 debug:</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep PTLDEBUG cfg/local.sh PTLDEBUG=${PTLDEBUG:--1} sles15build01:/home/hornc/lustre-filesystem/lustre/tests # ./llmount.sh ... Setting lustre.sys.jobid_var from disable to procname_uid Waiting 90 secs for update Updated after 4s: wanted 'procname_uid' got 'procname_uid' disable quota as required sles15build01:/home/hornc/lustre-filesystem/lustre/tests # lctl dk &gt; /tmp/dk.log </pre> </div></div> <p>Of interest are the messages with "TRACE" string. There are two types: lnet_parse() for incoming and lnet_handle_send() for outgoing:</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep TRACE /tmp/dk.log | grep -m 1 lnet_parse 00000400:00000200:19.0:1587582336.656655:0:12299:0:(lib-move.c:4290:lnet_parse()) TRACE: 0@lo(0@lo) &lt;- 0@lo : GET - for me sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep TRACE /tmp/dk.log | grep -m 1 lnet_handle_send 00000400:00000200:21.0:1587582336.657784:0:12301:0:(lib-move.c:1853:lnet_handle_send()) TRACE: 192.168.2.20@tcp(192.168.2.20@tcp:&lt;?&gt;) -&gt; 192.168.2.20@tcp(192.168.2.20@tcp:192.168.2.20@tcp) &lt;?&gt; : PUT try# 0 sles15build01:/home/hornc/lustre-filesystem/lustre/tests # </pre> </div></div> <p>Since this is a single node setup, all traffic should be sent via lolnd, but we can see that is not the case:</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep -c TRACE /tmp/dk.log 887 sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep TRACE /tmp/dk.log | grep -c 0@lo 4 sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep TRACE /tmp/dk.log | grep -vc 0@lo 883 sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep TRACE /tmp/dk.log | grep 0@lo 00000400:00000200:19.0:1587582336.656655:0:12299:0:(lib-move.c:4290:lnet_parse()) TRACE: 0@lo(0@lo) &lt;- 0@lo : GET - for me 00000400:00000200:19.0:1587582336.656681:0:12299:0:(lib-move.c:4290:lnet_parse()) TRACE: 0@lo(0@lo) &lt;- 0@lo : REPLY - for me 00000400:00000200:19.0:1587582336.656763:0:12299:0:(lib-move.c:4290:lnet_parse()) TRACE: 0@lo(0@lo) &lt;- 0@lo : PUT - for me 00000400:00000200:19.0:1587582336.656801:0:12299:0:(lib-move.c:4290:lnet_parse()) TRACE: 0@lo(0@lo) &lt;- 0@lo : ACK - for me sles15build01:/home/hornc/lustre-filesystem/lustre/tests # </pre> </div></div> <p>In fact only 4 messages were sent over lolnd.</p> <p>Doing the same thing with the fix we can see all messages use the lolnd:</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre>sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep -c TRACE /tmp/dk.log 462 sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep TRACE /tmp/dk.log | grep -c 0@lo 462 sles15build01:/home/hornc/lustre-filesystem/lustre/tests # grep TRACE /tmp/dk.log | grep -vc 0@lo 0 sles15build01:/home/hornc/lustre-filesystem/lustre/tests # </pre> </div></div> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/38312/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38312/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Introduce constant for the lolnd NID<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 56203e4ba0a64789e42ea45946e8c51f1db351fb</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/38313/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38313/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Primary NID of lolnd NID is the lolnd NID<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 33d2e44e5026f1e9162dd5e6b931085fdc035a34</p> <p>An exception to this is recovery messages. Recovery messages can not use the lolnd as that breaks the purpose of the attempt to ascertain the health of a local interface.</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/38388/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38388/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> ptlrpc: Check if NID is local, not just lolnd NID<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 95bcc24642c4b95d093407fef0947ee2f5a2c01a</p> <p>Landed for 2.14</p> <p><span class="error">&#91;Deleted&#93;</span></p> <p><span class="error">&#91;Deleted&#93;</span></p> <p><span class="error">&#91;Deleted&#93;</span></p> <p>Chris Horn (chris.horn@hpe.com) uploaded a new patch: <a href="https://review.whamcloud.com/38863" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38863</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Introduce constant for the lolnd NID<br/> Project: fs/lustre-release<br/> Branch: b2_12<br/> Current Patch Set: 1<br/> Commit: a0bb5d5f52414dcfc95bfa699c5d800c1d7daa7a</p> <p>Chris Horn (chris.horn@hpe.com) uploaded a new patch: <a href="https://review.whamcloud.com/38864" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38864</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Primary NID of lolnd NID is the lolnd NID<br/> Project: fs/lustre-release<br/> Branch: b2_12<br/> Current Patch Set: 1<br/> Commit: 7e924bc9a53ad82a0c4e6da4084e4de8536e4ada</p> <p>Chris Horn (chris.horn@hpe.com) uploaded a new patch: <a href="https://review.whamcloud.com/38865" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38865</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> ptlrpc: Check if NID is local, not just lolnd NID<br/> Project: fs/lustre-release<br/> Branch: b2_12<br/> Current Patch Set: 1<br/> Commit: 35c6cfeec9d5b9a354bc49319694d2360c48c68c</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/38863/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38863/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Introduce constant for the lolnd NID<br/> Project: fs/lustre-release<br/> Branch: b2_12<br/> Current Patch Set: <br/> Commit: b5fb6f1fe6dedca0ba6e405e90d7ec72d9b97e83</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/38864/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38864/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> lnet: Primary NID of lolnd NID is the lolnd NID<br/> Project: fs/lustre-release<br/> Branch: b2_12<br/> Current Patch Set: <br/> Commit: 4494f55aa2fe7d2ae274405fc0f2314193db1c0e</p> <p>Oleg Drokin (green@whamcloud.com) merged in patch <a href="https://review.whamcloud.com/38865/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/38865/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-12222" title="LNet should select loopback NI when possible" class="issue-link" data-issue-key="LU-12222"><del>LU-12222</del></a> ptlrpc: Check if NID is local, not just lolnd NID<br/> Project: fs/lustre-release<br/> Branch: b2_12<br/> Current Patch Set: <br/> Commit: 7635e6326de88bfd035c82ac9261e90e4f9b4a1b</p> Related Development Rank 1|i00fbz: Rank (Obsolete) 9223372036854775807