Thanks for your help. I will try another burn. I&#39;m using cdrecord and am using (I believe) good media. What command parameters for cdrecord would you recommend?<br><br>For fun I tried booting with a different external CD-ROM drive ... with identical results.
<br><br>I will also try the disk out on a R4600 Indy.<br><br>Thanks for your hard work.<br>MikeMartin<br><br><div><span class="gmail_quote">On 1/22/07, <b class="gmail_sendername">Kumba</b> &lt;<a href="mailto:kumba@gentoo.org">
kumba@gentoo.org</a>&gt; wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Mike Martin wrote:<br>&gt; I recently downloaded and attempted to boot an Octane with this disk. It
<br>&gt; died mounting the root partition:<br>&gt;<br>&gt; mount: Mounting /newroot/dev/loop0 on /newroot/mnt/livecd failed:<br>&gt; Invalid argument<br>&gt;<br>&gt; Not sure what happened. I assume I burnt the disk correctly else it
<br>&gt; wouldn&#39;t have made it that far. Any suggestions?<br>&gt;<br>&gt; MikeM<br><br><br>It&#39;s really hard to say.&nbsp;&nbsp;I tested it on all of my systems before uploading, and<br>&nbsp;&nbsp;with the Octane, this means an external drive (funny enough, an O2 CD drive
<br>jammed into a Sun 411 case).&nbsp;&nbsp;And that booted fine on both my Octane and Indy.<br><br>&quot;Invalid Argument&quot; from mount could mean a wide array of things (yay for Unix&#39;s<br>legacy of non-descriptive, ambiguous errors).&nbsp;&nbsp;The process that occurs on an SGI
<br>bootcd for us is a rather complex one:<br><br>1. arcload boots from the DVh partition of the CD<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;(yes, these CDs have partitions)<br>2. arcload finds and boots a kernel<br>3. kernel loads, and executes /init in an embedded initramfs file linked
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;into the kernel<br>4. /init does some prep work, and launches `getdvhoff` to scan the CD<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;for the offset of the next partition (where / lives), and passes a<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;number representing this offset back to `losetup`.
<br>5. losetup uses this number to &quot;point&quot; /dev/loop0 at this offset, which<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;effectively makes /dev/loop0 a block device with data on it.<br>6. mount tries to mount /dev/loop0 and pivot_root into the real Gentoo
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;filesystem.<br><br><br>Quite likely, step #5 might&#39;ve failed somewheres along the line.&nbsp;&nbsp;The offset has<br>to be exact to the bit, so maybe something got whacked in the burn and the<br>detected offset is invalid.&nbsp;&nbsp;Hard to say without more information.&nbsp;&nbsp;Thus, when
<br>it got to step #6, boom.<br><br>I&#39;d try re-burning the disk at a slower speed, use only CD-R&#39;s of decent quality<br>(TDK, Memorex, Sony, Ricoh/Ritek, etc,.. brands), and use cdrecord (or whatever<br>license-unencumbered version is out there.&nbsp;&nbsp;stupid license wars).&nbsp;&nbsp;A few people
<br>reported getting it to work with a windows burn tool, but we have little data on<br>that, thus why cdrecord is the suggested tool.<br><br>Mostly, you were able to read the kernel into memory, which is ~8MB.&nbsp;&nbsp;It&#39;s<br>
possible the disc you burned was good enough to get those 8MB off to boot the<br>kernel, but when it went looking for the meat, it got denied and pwned.<br><br><br><br>--Kumba<br><br>--<br>Gentoo/MIPS Team Lead<br><br>&quot;Such is oft the course of deeds that move the wheels of the world: small hands
<br>do them because they must, while the eyes of the great are elsewhere.&quot;&nbsp;&nbsp;--Elrond<br>--<br><a href="mailto:gentoo-mips@gentoo.org">gentoo-mips@gentoo.org</a> mailing list<br><br></blockquote></div><br>