How can it read an optical for boot & then not want to finish setup from it?
 

Bought a new bare-bones kit from Tiger Direct & a new copy of Slack 13,37 (Linux.2.6.37.6) from Slack Store: Assembled this kit:
MSI H61M-P23 MB (sandybridge/b3 stepping)
Intel i3 cpu
4GB memory
Seagate Barracuda 0.5 TB SATA HDD
Sony relabled - LiteOn SATA DVD-RW (in Sony OEM packging) my HW BIOS reports this as an Otiarc DVD RW AD-726

Note: Beware of this bit of "bate-and-switch" by TD. Pre-Purchase, the HCL indicated to me that the *advertised* LiteOn ODD would likely work, the Sony is a different matter. I let TD know I wasn't happy with the switch, but have had no response so far. But this is not my concern here, just information.

Post Assembly:
Put the DVD in the drive, set the bios/boot to it and hit reset. Went right into the boot all the way into setup. At this point setup can't find the source media - fail.

Using a usb Toshiba Super-Multi-Drive, also a dvd-RW, I put the dist. dvd in it, reset the boot/bios and hit reset. loaded *everything* into four ext2 partitions: sda1-swap, 2-/root, 3-/home, 4- /usr/local (this system has never even sniffed windows os media)

shut it down. removed the toshiba usb dvd, put the media back into the internal dvd, re-reset the boot bios and hit reset....

system boots up to where I input "huge.s /dev/sda2 rdinit= ro" line in, system starts & alls good except:

Know Prpblems:
1) mount tells me that the internal dvd is not mounted.
2) ifconfig -a only responds with the local loopback, although I *asked* setup to set up the network during installation. - note that the wire is a known good premise-routed dsl connection, modem in bridge mode to my router using dhcp. Connection brainlessly easy on a Win7 platform with no special user/pw connection issues or input, that data being in the router configuration, already.

ancillary discussion & comments:
1. the network issue has not been worked beyond ifconfig, so I'm not askiing questions about that here, haven't done enough to investigate. just reporting known status

2. what does the "rdinit=" piece of the kernel boot command (above) do? Isn't it set to null? If I leave it out, it goes on into the boot-kernel/setup-routine off the dvd (and fails as described above), instead of the current installation on /dev/sda2. The "ro' is a separate field indicating the root file system is to be started read only for fsck considerations, isn't it?

3. my suspicion is that it won't load the media from the internal dvd because it's already discarded it as unsupported - doing that after initial boot segment, running at-that-point on the setup kernel. But since it CAN/DID read it, why can't it GO ON and read it in whatever mode or configuration it had to begin with? Isn't there a "generic dvd driver" being used in those initial moments? Anyone know what really goes on there?

4. Would anyone know the "what/why" on an interesting artifact seen while doing the fdisk during sw installation above? To whit:
fdisk reported the "advertised" disk capacity (.5TB), but there were 2048 sectors that seemed to be *reserved*, that is, they could not be dedicated to any partition. fdisk working on thes blank disk, supplied the first allocateable sector as 2048, which is where sda1 begins.

I can't say for sure but many dvd/cd live and install disks do not show up as a mount as a normal dvd would on a hard drive running system.

You need to fix the nic. That is common. I'd try a few other live cd's.

I can't say if this distro is any good or any way to test it for quality.

Should have started with memtest for few days.


I have had only good results from tigerdirect. If you were unhappy then you should have returned it.

Thanks!
I too have always had good experiences with TD, I think this is the 4th I've purchased, all good, it's just that the ODD is a relabled version of the advertised hdwe. If I was running MS/Wndows it would probably be a non-issue. If TD had responded to my email to their customer service, I likely would have returned it, but at this point its nearly becomming moot. They probably have a disclaimer somewhere in the fine print for substitutioins anyway.

I'll insert a different cd and try to mount it. good test idea, thanks. I believe the *factory* Slackware DVD is good, as it was able to be used in building the system as it exists, which seems functional in all aspects exept as noted in the original notes.

I did run memtest for *several* hours, and a full hard-core (testing) format during installation. I was getting rabid to see it run...

I have to do a lot more reading before I can address the nic. I have no reason to believe it faulty, the board is a pretty new model, so I figured I might find a few hdwe/software concurrency issues. All I could do is use the TD discriptions for checking the HCL, which seemed OK.

So I tried to see if different media would make a differnce (see Jefro note above.) ...I
burned a dvd (on a different system) with a firefox tarball and attempted to mount after insertion.
I didn't reboot after insertion, since the internal ODD wasn't mounted, I did't think taking the distribution cd out would give the system a problem....

tried the command:

mount -t iso9660 /dev/sdb . (also tried with /dev/sdb1) ("." being my home (root) directory)

system response was

mount: special device /dev/sdb does not exist

which I guess means that it still isn't looking at the internal ODD

Second test (attempted): put the tarball disk back into the usb DVD, put the distribution disk back into the internal ODD, rebooted
At which point I realized I don't know where the USB is in the filestructure or how to mount or where that kind of device is nominally mounted.
since that USB dvd drive was used to build the system as it exists, I would have thought it would have mounted it when it found it during boot, but mount has no record.......
I'm going to go search dmesg for a clue where it was, hopefully that hasn't been overwritten

The mechanical experiments are great, I have to assume some knowledge comes from each one, but I'd really like to focus on the original question, how can an ODD be a "now you see it, now you don't" - seems a flaw.

If if didn't see it enough to boot from it, I think I'd understand... but this thing dissapears as far as the kernel is concerned, shouldn't the kernel hold onto some kind of low level cognizance?

On my system (Fedora) the dvd drive is /dev/sr0; both /dev/cdrw and /dev/cdrom are symlinks to /dev/sr0. You might try mounting one of those devices, e.g. (I use /opt as a convenient mount point because Fedora doesn't place anything in that directory by default)

I would suggest setting the BIOS settings for the SATA controller to IDE compatability and turn any RAID mode off. Also as suggested try another distribution. It appears that Ubuntu works.

Not an expert but because of the BIOS/boot process it is possible to boot with an OS/installer that does not support the controller. When the installer tries to mount the media it fails because the necessary module(s) are not included or complied into the kernel.

http://www.linuxquestions.org/questi...tected-893254/

After beating my head against it, I thought I just go and get another ODD. So I read the HCL (a paucity of SATA ODDs in there) and took my best shot (Pioneer, also the cheapest available....) same problem. So that new one came back out, and I started to hook up the original.

Now don't ask me why I decided to do it, but I took a notion to move the ODD to SATA 3 connector (they were 1 & 2: HDD & ODD, now 1&3) and Bingo It workits!

Did a complete new build & config....(wanted to work that dog).

Mother board docs doen't make an notion towards setting up like that, but oh well. The very first assembly was with the ODD on SATA1 and the HDD on SATA2, but the problem was first seen in that configuration, so I switched to the HDD on 1 and the ODD on 2, only to have the same problem.. SO I wasn't originally suspecting the port hardware. I suspect it's something like RockDoctor was expressing above.

Thank the SlackGods that the ODD turned out to be supported

Lots of questions remaining, but for now I'll abandon them in pursuit of the network setup that still doesn't want to play.
Ping of the "lo" is successful, so it's really got to be just some kind of configuration element I'm missing.

REALLY APPRECIATE ALL THE GREAT SUGGESTIONS! each provides another insite vantage to consider

Id still like to find out what the rdinit= piece of the kernel boot instruction is....
but for now I am happy.

I want to thank Michaelk for the link to the other guy who was also having a nic problem on the same board. thats on a ubuntu dist, but somehow it will be applicable to slackw

O my system the odd's are 3 & 4

eddy1,
I'm still good here with it on 3. It would be interesting as hell to find out what really goes on there, but stopping at success is a logical root for now.... maybe someday we'll know why.
You'd think the MB docs would say something; URLs pointing at the details, maybe ( ideally ? ), but that's the glory of being an anit-Jobian... lots of choices, and there are probably too many for the MB mfg to consider or keep track of.

Thanks for the data !

-3dB

Well my asus MB stated it somewhere in manual & I skimmed over it & caught it, so maybe you did he sam & missed it

Eddy1,
Thanks for your response.

I looked at the setup booklet again - nothing. Just locations pointed out for assembing the system wires, with few other comments or notes.