1.5M Hour MTBF For Boot Drive

[HarryMuscle]

<div class="bbWrapper">Just curious if the professionals here (ie: those that run TrueNAS for work or serious home use, not your average hobby use) would consider a pair of 1.5 million hour MTBF rated SSD drives to be suitable for the boot drive (Transcend 800S series)? Everything else in the server is rated at 2.5 million (HGST drives) or 3 million (Micron SSDs) hours. I know MTBF is not the greatest metric but it does give a general idea of the reliability of a drive when averaged and professionals usually (or at least they should) care about how reliable the final product will be to minimize down time.<br /> <br /> Thanks,<br /> Harry</div>

 

[ChrisRJ]

<div class="bbWrapper">I am personally not overly concerned with the boot drive. Worst case (at least for Core) is to re-install and restore the configuration. That should be less than 30 minutes and a rather simple process.</div>

 

[DigitalMinimalist]

<div class="bbWrapper">To add:<br /> Boot drive size &amp; speed and reliability don’t really matter.<br /> Just backup the TrueNAS config regularly</div>

 

[Etorix]

<div class="bbWrapper">1.5 M hours is a little over 171 <b>years</b>. As a non-professional, I'd be quite happy if the boot drive were to follow this metric and be the first component to wear out.</div>

 

[danb35]

<div class="bbWrapper">MTBF is pretty much a meaningless metric. If 150,000 of the devices lasted 10 hours without failure, there's your 1.5Mhour MTBF.</div>

 

[HarryMuscle]

<div class="bbWrapper"><blockquote data-attributes="member: 106202" data-quote="Etorix" data-source="post: 711609" class="bbCodeBlock bbCodeBlock--expandable bbCodeBlock--quote js-expandWatch"> <div class="bbCodeBlock-title"> <a href="/community/goto/post?id=711609" class="bbCodeBlock-sourceJump" rel="nofollow" data-xf-click="attribution" data-content-selector="#post-711609">Etorix said:</a> </div> <div class="bbCodeBlock-content"> <div class="bbCodeBlock-expandContent js-expandContent "> 1.5 M hours is a little over 171 <b>years</b>. As a non-professional, I'd be quite happy if the boot drive were to follow this metric and be the first component to wear out. </div> <div class="bbCodeBlock-expandLink js-expandLink"><a role="button" tabindex="0">Click to expand...</a></div> </div> </blockquote>MTBF doesn't actually work that way. It's the amount of hours a certain amount of drives (usually several thousand) work between or before a failure. So for example 1000 of these drives would work for 1500 hours before or between failures.<br /> <br /> Thanks,<br /> Harry</div>

 

[Etorix]

<div class="bbWrapper"><blockquote data-attributes="member: 112058" data-quote="HarryMuscle" data-source="post: 711713" class="bbCodeBlock bbCodeBlock--expandable bbCodeBlock--quote js-expandWatch"> <div class="bbCodeBlock-title"> <a href="/community/goto/post?id=711713" class="bbCodeBlock-sourceJump" rel="nofollow" data-xf-click="attribution" data-content-selector="#post-711713">HarryMuscle said:</a> </div> <div class="bbCodeBlock-content"> <div class="bbCodeBlock-expandContent js-expandContent "> MTBF doesn't actually work that way. </div> <div class="bbCodeBlock-expandLink js-expandLink"><a role="button" tabindex="0">Click to expand...</a></div> </div> </blockquote>It doesn't work the way you state either…<br /> It has to be statistical measured (or estimated) over a number of drives, but the resulting figure applies to a <i>single</i> drive. Else, one could just take one million drives and be happy that they fail after a mere one hour and a half…<br /> Some maths here: <a href="https://en.wikipedia.org/wiki/Mean_time_between_failures" target="_blank" class="link link--external" rel="nofollow ugc noopener">https://en.wikipedia.org/wiki/Mean_time_between_failures</a><br /> <br /> A MTBF (or rather MTTF in this case) of 1.5e6 hours obviously doesn't guarantee that the drive will only fail after 171 years. But the latter figure may be a more telling indication of how reliable it is.<br /> Or, to throw in another figure, it works out to an Annualised Failure Rate of 0.58%. Is that low enough?<br /> <br /> <blockquote data-attributes="member: 112058" data-quote="HarryMuscle" data-source="post: 711590" class="bbCodeBlock bbCodeBlock--expandable bbCodeBlock--quote js-expandWatch"> <div class="bbCodeBlock-title"> <a href="/community/goto/post?id=711590" class="bbCodeBlock-sourceJump" rel="nofollow" data-xf-click="attribution" data-content-selector="#post-711590">HarryMuscle said:</a> </div> <div class="bbCodeBlock-content"> <div class="bbCodeBlock-expandContent js-expandContent "> professionals usually (or at least they should) care about how reliable the final product will be to minimize down time. </div> <div class="bbCodeBlock-expandLink js-expandLink"><a role="button" tabindex="0">Click to expand...</a></div> </div> </blockquote>If down time is the concern, I further note that you have a <i>pair</i> of boot drives. If these can be hot-swapped, there's no down time involved to replace a single failing drive; it would take simultaneous failure by both drives to have to reinstall and restore a configuration file—half an hour down time, as pointed to by <a href="https://www.truenas.com/community/members/104866/" class="username" data-xf-init="member-tooltip" data-user-id="104866" data-username="@ChrisRJ">@ChrisRJ</a>.</div>