<div dir="auto">No you are not asking that much. <div dir="auto"><br></div><div dir="auto">Build your own! </div><div dir="auto"><br></div><div dir="auto">Thomas</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Sat, Sep 25, 2021, 2:56 PM <a href="mailto:paulz@ieee.org">paulz@ieee.org</a> <<a href="mailto:paulz@ieee.org">paulz@ieee.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div style="font-family:Helvetica Neue,Helvetica,Arial,sans-serif;font-size:13px"><div></div>
<div dir="ltr">So when I am getting the NAS box, I want at least 2 bays.</div><div dir="ltr"><br></div><div dir="ltr">It comes down to how much actual storage I need and how much it costs. Raid 5 will require a more expensive box and several more disk drives. So it is not obvious which will give the best total price per bit.</div><div dir="ltr"><br></div><div dir="ltr">And I probably want to steer clear of brands touting that they have all of the fancy software installed. A reviewer of one model complained that after he bought NAS to have on his own network, the only way to access his personal data was through the website of WD.</div><div dir="ltr"><br></div><div dir="ltr">In my definition, my NAS should appear to my computer as a drive that can be accesses as easily as the internal drive, after things are set up. This independent of the OS on the computer. Am I asking too much?</div><div dir="ltr"><br></div><div dir="ltr"><br></div><div><br></div>
</div><div id="m_4939435283143834309ydpc691be4byahoo_quoted_2619183421">
<div style="font-family:'Helvetica Neue',Helvetica,Arial,sans-serif;font-size:13px;color:#26282a">
<div>
On Saturday, September 25, 2021, 01:41:10 PM PDT, Rick Moen <<a href="mailto:rick@linuxmafia.com" target="_blank" rel="noreferrer">rick@linuxmafia.com</a>> wrote:
</div>
<div><br></div>
<div><br></div>
<div>Quoting Paul Zander (<a shape="rect" href="mailto:paulz@ieee.org" rel="nofollow noreferrer" target="_blank">paulz@ieee.org</a>):<br clear="none"><br clear="none">> Raid 1 uses 2 drives with all data written to both. Protection of<br clear="none">> loss of 1 drive, but only get storage area of 1 drive. For RAID to be<br clear="none">> "efficient" 4 or more drives are needed. Did I get that right?<br clear="none"><br clear="none">Head of the class, sir! (Technically, RAID5 requires at least _3_<br clear="none">drives.)<br clear="none"><br clear="none">The significant advantage of RAID1 mirroring over RAID5 & similar is <br clear="none">simplicity and lower computational overhead (especially, but not<br clear="none">exclusively, during a rebuild cycle following failure and replacement <br clear="none">of a drive). So, in many scenarios, those RAID1 advantages are<br clear="none">compelling and make the 50% "loss" of capacity of little consequence and<br clear="none">easy to justify.<br clear="none"><br clear="none">With (say) RAID5, you have around "loss" of capacity equal to the <br clear="none">capacity of one of the constituent drives. <br clear="none"><br clear="none">Here is a FAQ: <a shape="rect" href="https://www.vantagetech.com/faq/raid-5-recovery-faq.html" rel="nofollow noreferrer" target="_blank">https://www.vantagetech.com/faq/raid-5-recovery-faq.html</a><br clear="none"><br clear="none"><br clear="none">I alluded earlier to ZFS (and its slightly lackluster Linux-native<br clear="none">imitator, btrfs) as being in a different category relative to regular<br clear="none">*ix filesystems like ext4, in ways that make it much more desirable for<br clear="none">servers. I should elaborate.<br clear="none"><br clear="none">ext4 is an excellent, high-performing, conservatively designed *ix<br clear="none">filesystem. There is nothing wrong with it. Things it doesn't do, and<br clear="none">doesn't aspire to do, include:<br clear="none"><br clear="none">o background fsck and auto-repair / self-healing during normal system operation<br clear="none">o checksumming and vetting of every byte written, data and metadata<br clear="none">o data snaphots and replication<br clear="none">o native volume management<br clear="none">o native handling of RAID<br clear="none">o automatic rollback, in the event of detected error or inconsistency<br clear="none">o native data compression and de-duplication<br clear="none">o a lot more. see <a shape="rect" href="https://en.wikipedia.org/wiki/ZFS#Summary" rel="nofollow noreferrer" target="_blank">https://en.wikipedia.org/wiki/ZFS#Summary</a><br clear="none"><br clear="none">A regular *ix filesystem, even a really good one like ext4, has none of<br clear="none">those things built-in. As filesystem sizes and data collections <br clear="none">(e.g., A/V files) get bigger and bigger, the risks from bitrot and <br clear="none">data corruption silently occurring and accumulating over time become<br clear="none">more worrisome, and also the risk of having multiple days of system<br clear="none">downtime just because of need to correct filesystem errors during a<br clear="none">reboot-driven fsck -- because you had 10 terabytes of files.<br clear="none"> <br clear="none">The whole idea of a file _server_ on your network is supposed to be for<br clear="none">it to be a _reliable_ place to house your files, which is where the<br clear="none">first two ZFS advantages listed really shine. On the minus side, ZFS is<br clear="none">fairly RAM-thirsty. So, arguably worth the RAM overhead.<br clear="none"><br clear="none">ZFS was one of several ground-breaking Sun Microsystems projects that<br clear="none">gave Solaris and several compatibly licensed BSDs a functionality edge<br clear="none">over Linux. The others were dtrace (which is kernel-level debugging),<br clear="none">Solaris Containers aka Zones (which are a better implementation of<br clear="none">chroot jails), a Kernel-based Virtual Machine, and OpenSolaris Network<br clear="none">Virtualization and Resource Control aka Crossbow (a set of features that<br clear="none">provides an internal network virtualization and quality of service).<br clear="none">All of those are C-coded in the Solaris kernel under CDDL terms, which<br clear="none">clashes with Linux's GPLv2 terms, hence the code, even if ported to the<br clear="none">Linux kernel, cannot be lawfully distributed in binary form, as that<br clear="none">would be copyright violation.[1] <br clear="none"><br clear="none">btrfs is an independent attempt to implement most of ZFS's features in<br clear="none">Linux. However, even now, 14 years after its debut, it still has some<br clear="none">problems. Here, this runthrough will tell you about that, better than I<br clear="none">could:<br clear="none"><a shape="rect" href="https://arstechnica.com/gadgets/2021/09/examining-btrfs-linuxs-perpetually-half-finished-filesystem/" rel="nofollow noreferrer" target="_blank">https://arstechnica.com/gadgets/2021/09/examining-btrfs-linuxs-perpetually-half-finished-filesystem/</a><div id="m_4939435283143834309ydpc691be4byqtfd68902"><br clear="none"><br clear="none"><br clear="none"><br clear="none">_______________________________________________<br clear="none">conspire mailing list<br clear="none"><a shape="rect" href="mailto:conspire@linuxmafia.com" rel="nofollow noreferrer" target="_blank">conspire@linuxmafia.com</a><br clear="none"><a shape="rect" href="http://linuxmafia.com/mailman/listinfo/conspire" rel="nofollow noreferrer" target="_blank">http://linuxmafia.com/mailman/listinfo/conspire</a><br clear="none"></div></div>
</div>
</div></div>_______________________________________________<br>
conspire mailing list<br>
<a href="mailto:conspire@linuxmafia.com" target="_blank" rel="noreferrer">conspire@linuxmafia.com</a><br>
<a href="http://linuxmafia.com/mailman/listinfo/conspire" rel="noreferrer noreferrer" target="_blank">http://linuxmafia.com/mailman/listinfo/conspire</a><br>
</blockquote></div>