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/LUDOC-175 Lustre Documentation <p>It would help customers and partners if we could help add information in the following areas:</p> <p>1. More info on the parameters<br/> 2. Suggestions around selecting the range of the parameters</p> <p>Thanks</p> LUDOC-175

Need Guidance on Selecting sgpdd-survey Parameters

Improvement Major Resolved Low Priority Dan Cobb Dan Ferber Thu, 22 Aug 2013 18:09:46 +0000 Wed, 25 Oct 2017 15:41:09 +0000 Wed, 25 Oct 2017 15:41:09 +0000 0 2 <p>(from Zhiqi Tao)</p> <p>These failures are normal. There should be a sgpdd_survey_<span class="error">&#91;date&#93;</span>@<span class="error">&#91;time&#93;</span>.detail file in the working directory that shows the failures due to out of memory.</p> <p>"sg starting in command at "sgp_dd.c":827: Cannot allocate memory"</p> <p>BTW, I normally put sgpdd-survey into a script and run from there.</p> <p>Here is what I normally use.</p> <p>$ cat run_sgpdd-survey.sh<br/> #!/bin/bash</p> <ol> <li>Device List, for example, "/dev/raw/raw1 /dev/raw/raw2 /dev/raw/raw3"<br/> DEVICE="regal-oss01:/dev/sdh regal-oss01:/dev/sdi regal-oss01:/dev/sdj regal-oss01:/dev/sdk regal-oss01:/dev/sdl regal-oss01:/dev/sdm regal-oss00:/dev/sdh regal-oss00:/dev/sdi regal-oss00:/dev/sdj regal-oss00:/dev/sdk regal-oss00:/dev/sdl regal-oss00:/dev/sdm"</li> </ol> <ol> <li>The directory for the test output<br/> OUTDIR="/root/test_results/sgpdd_survey/"</li> </ol> <ol> <li>The test dataset size (MB) for each LUN. The total dataset size must be larger than 2 times of the RAM size in order to avoid the caching.</li> <li>The calculation of the size</li> <li>(RAM size * 2) / the number of LUNs</li> <li>For example, the server RAM is 24GB and there are 5 LUN</li> <li>the size should be (24GB * 2 ) / 5 ~= 10 GB = 10240 MB</li> <li>However, it is a good idea to run a very short trial run to make sure the test configuration working properly before scheduling a complete test.</li> <li>SIZE=100 is a good number for the trial run</li> </ol> <ol> <li>SIZE=100<br/> SIZE=23000</li> </ol> <ol> <li>Concurrent Regions</li> <li>CRGLO=1 - start from 1 concurrent regions</li> <li>CRGHI=256 - finish at the X number of concurrent regions.</li> <li>Before starting a complete run, use the following settings for a quick trial run.<br/> CRGLO=1<br/> CRGHI=256</li> </ol> <ol> <li>Thread count</li> <li>THRLO=1</li> <li>THRHI=4096</li> <li>Before starting a complete run, use the following settings for a quick trial run.<br/> THRLO=1<br/> THRHI=4096</li> </ol> <p>mkdir -p ${OUTDIR}<br/> ssh regal-oss00 mkdir -p ${OUTDIR}<br/> echo crglo=$CRGLO crghi=$CRGHI thrlo=$THRLO thrhi=$THRHI size=$SIZE rslt_loc=${OUTDIR} scsidevs="$DEVICE" ./sgpdd-survey<br/> crglo=$CRGLO crghi=$CRGHI thrlo=$THRLO thrhi=$THRHI size=$SIZE rslt_loc=${OUTDIR} scsidevs="$DEVICE" ./sgpdd-survey</p> <p>I used the enhanced sgpdd-survey from <a href="https://jira.hpdd.intel.com/browse/LU-2043" class="external-link" target="_blank" rel="nofollow noopener">https://jira.hpdd.intel.com/browse/LU-2043</a> According to the ticket, It should be in 2.4 now.</p> <p>I received the same error... The root cause is the sg device. If you try to run sgp-dd against the sd device, you don't receive any errors, but if you try (as sgpddsurvey does) the sg device you receive the memory error. My workaround (per Zhiqi suggestion) was using raw devices.</p> <p>Gabriele Paciucci</p> Development Rank 1|hzvyzb: Rank (Obsolete) 9877