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-8483 Lustre <p>Running sanity.sh on 1 client, 1 MDT, and 1 OST, I consistently find it fails just after test 117 due to a get_params error:</p> <div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent"> <pre> SKIP: sanity test_117 skipping ALWAYS excluded test 117 error: get_param: param_path 'osc/lustrewt-OST*-osc-*/resend_count': No such file or directory error: set_param: param_path 'osc/lustrewt-OST*-osc-*/resend_count': No such file or directory sanity.sh returned 0 Finished at Sun Aug 7 12:05:37 PDT 2016 in 2657s ./auster: completed with rc 0 </pre> </div></div> <p>Even if test 117 is excluded (as above), it still aborts. It tells me that auster completed with rc of 0, but that is not true. A test earlier did fail, and there are more tests which did not get run because of this abort.</p> LU-8483

sanity.sh aborts after test 117 due to get_param error

Bug Major Resolved Fixed Doug Oucharek Doug Oucharek Sun, 7 Aug 2016 19:13:30 +0000 Mon, 8 Aug 2016 21:39:28 +0000 Mon, 8 Aug 2016 21:39:28 +0000 Lustre 2.9.0 0 2 <p>This error message is generated by <tt>set_resend_count()</tt> which is run between <tt>test_117()</tt> and <tt>test_118a()</tt>. It appears that the client is not mounted at this point for some reason, so some kind of console and/or test logs are needed to see what test is actually failing.</p> <p>Turns out the problem is this:</p> <ul class="alternate" type="square"> <li>We are testing Multi-Rail, so the interfaces are known by multiple NIDs.</li> <li>Autotest uses DNS names (I've tried IP addresses and ran into some problems so reverted to DNS names).</li> <li>In our testing, we mounted the file system using a NID of the MDT which did not match the DNS name. That worked, but caused this problem later on.</li> </ul> <p>So, the lesson is this: when Multi-Rail comes into play, we must make sure that the DNS names for the nodes refer to the primary NID (preferably) or at least is the NID which we are mounting the file system with.</p> Development Rank 1|hzyjp3: Rank (Obsolete) 9223372036854775807 Severity