<div class="gmail_quote">On Wed, Feb 23, 2011 at 10:28 PM, <a href="mailto:general_email@technicalbloke.com">general_email@technicalbloke.com</a> <span dir="ltr"><<a href="mailto:general_email@technicalbloke.com">general_email@technicalbloke.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">
<div class="im">On 23/02/11 15:52, Andrew Farnsworth wrote:<br>> When it comes to HDD, the best thing is to just replace it. Costs are so<br>> low today that if there is any question at all about the health of a HDD it<br>
> is better to pull that drive and replace it with a new one. Either<br>> destroy/discard the questionable drive or move it to a non-critical use.<br>><br>> Testing the MB etc, there are several bootable "live CD" images with many<br>
> tools for testing all kinds of Hardware. While this may not suit your needs<br>> (i.e. the computer in question is in a data center or something) you can<br>> certainly get a good idea of what tools are current by what is on the latest<br>
> Live CDs for testing.<br>><br>> Andy<br>><br>> On Wed, Feb 23, 2011 at 10:41 AM, Dylan <<a href="mailto:dylan@dylan.me.uk">dylan@dylan.me.uk</a>> wrote:<br>><br>>> Hi All,<br>>><br>>> Any recomendations for a h/ware diagnostics boot disk? I'm looking for<br>
>> tests<br>>> of HDD and as much of the motherboard as can be tested in software ...<br>>><br>>> Any pointers gratefully received!<br>>><br>>> Dx<br><br><br></div>Andrew, pls don't top post, and also there's no need to be so profligate<br>
with somebody else's money! It's absolutely worth checking your disks<br>for errors otherwise you're replacing them on pure superstition alone!</blockquote>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid"><br>I generally use seatools although it doesn't always run properly on all<br>hardware (based on crappy FreeDOS rather than linux sadly). All the<br>
drive manufacturers have their own free (beer) testing tools and if you<br>want to send a drive back under warranty they will insist you use their<br>tool to verify it is bad before giving you an RMA.<br><br>Smartctl (part of smartmontools), while not a boot disk, is available<br>
on/via many boot disks and will give you a lot of info about your drive<br>and it's SMART data. Be careful though, I've just read (in the link to<br>stresslinux from John Hearn's post) that it can cause bad blocks on<br>
Samsung F4 EcoGreen drives which might explain why the FreeNAS I built<br>last year has already chewed it's way through 3 of the damn things!!!<br><br>Roger</blockquote>
<div> </div>
<div>Roger et al.</div>
<div> I appologize for top posting earlier. As to replacing the disk, I work daily with data that is much more valuable than the underlying disks themselves. Therefore, I answered on the assumption that the disk is the least expensive part of the issue. From a personal perspective, I certainly agree that just throwing money at it is not the correct first option. As to Smartctrl, it works very well for accessing the SMART functionality of drives, however, it has been shown that SMART does not catch many of the issues that are killing drives until they are already dead / dying. Using the manufacturer's test suite is usually best for getting the manufacturer to fix/replace a drive under warranty but be aware that some of their "tests" can be destructive in nature.</div>
<div> </div>
<div>The Ultimate Boot CD and the System Rescue CD both come with several tools for checking drives as well as recovering data from them.</div>
<div> </div>
<div> I would emphasise again that if the data is really more valuable than the disk it is sitting on, it is worth spending a bit of money to be sure it is safe. Either via backups (onsite or off) and/or via redundance (i.e RAID or even just multiple copies on multiple drives).</div>
<div> </div>
<div>I personally have a set of drives where I keep throw away data and I run these in RAID 0 mode (striped with no redundancy) so that I have amazing throughput but don't care if I lose everything if a drive fails. I then have a set of drives I run in RAID 1 mode (mirrored with single drive redundance) and everything on my Mirrored array is backed up offsite to cover me if there is a localized disaster.</div>
<div> </div>
<div>Ok, discovered on rereading that I have gone a bit off topic. Getting back on topic, if a tool reports a few bad blocks, that is normal, if you start getting a lot of them, get everything off that drive as soon as possible.</div>
<div> </div>
<div>Andy</div></div>