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-14164 Lustre <p>On my system (for some reason) resolving an unknown hostname returns the loopback IP (127.0.0.1).</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> &gt;# getent hosts mgs.exports.192.168.122.100 ::1 mgs.exports.192.168.122.100.localhost </pre> </div></div> <p>this ends up breaking specific commands, like</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> lctl get_param -n *.MGS*.exports.192.168.122.100@tcp.uuid</pre> </div></div> <p>this ends up trying to look at:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> mgs/mgs/exports/192/168/122/100@tcp999/uuid</pre> </div></div> <p>The reason is that get <tt>hostbyname()</tt> called from <tt>clean_path()-&gt;libcfs_str2nid()</tt> returns <tt>127.0.0.1</tt> for "<tt>exports.192.168.122.100"</tt>. This leads the to the NID not found in the string which leads to the IP address being included in the path.</p> <p>To be clear this seems to be an issue on my setup. Filing this ticket to raise awareness that this could happen.</p> <p>I ran into it while trying to run sanity-sec:test_31</p> LU-14164

libcfs_str2nid() gives undesirable output

Bug Minor Open Unresolved WC Triage Amir Shehata Tue, 1 Dec 2020 02:15:18 +0000 Tue, 1 Dec 2020 18:43:35 +0000 0 2 <p>I think it would be ok to return <tt>LNET_NID_ANY</tt> in case of 127.0.0.1, since we already do not allow the TCP loop back interface to be configured for LNet. That is typically an error due to "$hostname" being aliased to <tt>localhost</tt>, and using the LNet <tt>0@lo</tt> interface is more efficient. </p> <p>See, for example, the code in <tt>lustre/utils/mkfs_lustre.c</tt> that verifies the server NID:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> nid = libcfs_str2nid(s1); <span class="code-keyword">if</span> (nid == LNET_NID_ANY) { fprintf(stderr, <span class="code-quote">"%s: Cannot resolve hostname <span class="code-quote">'%s'</span>.\n"</span>, progname, s1); free(converted); <span class="code-keyword">return</span> NULL; } <span class="code-keyword">if</span> (strncmp(libcfs_nid2str(nid), <span class="code-quote">"127.0.0.1"</span>, strlen(<span class="code-quote">"127.0.0.1"</span>)) == 0) { fprintf(stderr, <span class="code-quote">"%s: The NID <span class="code-quote">'%s'</span> resolves to the loopback address <span class="code-quote">'%s'</span>. Lustre requires a non-loopback address.\n"</span>, progname, s1, libcfs_nid2str(nid)); </pre> </div></div> <p>It may be that returning <tt>LNET_NID_ANY</tt> for 127.0.0.1 would prevent this message from being printed properly, which I think would be confusing to users, so it may be that <tt>clean_path()</tt> needs a similar check:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> nid = libcfs_str2nid(tmp); <span class="code-keyword">if</span> (nid != LNET_NID_ANY) { nidstr = libcfs_nid2str(nid); <span class="code-keyword">if</span> (!nidstr) <span class="code-keyword">return</span> -EINVAL; + /* some DNS resolvers <span class="code-keyword">return</span> localhost <span class="code-keyword">for</span> unknown hosts, + * but <span class="code-keyword">this</span> can never be valid NID <span class="code-keyword">for</span> any mounted client. + */ + <span class="code-keyword">if</span> (strncmp(nidstr, <span class="code-quote">"127.0.0.1"</span>, strlen(<span class="code-quote">"127.0.0.1"</span>)) == 0) { + nidstr = NULL; + <span class="code-keyword">continue</span>; + } <span class="code-keyword">break</span>; } </pre> </div></div> <p>Amir, can you see if this solves your problem?</p> Development Rank 1|i01g1j: Rank (Obsolete) 9223372036854775807 Severity