Install Problem - core 3.6 - kernel 2.6.33-3 tiny

[Chan]

Hello,
I've been having a problem installing microcore 3.6 - linux kernel 2.6.33.3 tinycore.

Computer I am using is an Compaq Presario S4020WM with a 40Gb internal drive and Cdrom. Both are via IDE interface. Computer has 1Gb of ram, Athlon XP 2400+ processor (2000Mhz/266Mhz). Full specs are here -

 "h10025.www1.hp.com/ewfrf/wc/document?docname=c00021854&lc=en&dlc=en&cc=us&lang=en&product=326832"

The image I downloaded and burned to a CD was tinycore-current. Prior to booting the above PC with this CD I booted to an old 6.2 DOS disk and deleted all drive partitions with Fdisk and also did an Fdisk /mbr

When booting tinycore-current CD all seemed normal up to the point right after the message Setting Timezone.  Right after that I get several hdd: media errors - as can be seen on the attached picture.  I've tried using 2 different cdrom drives (both known to be good) and get the same error.

Any help or insight would be appreciated.

Thanks

[gerald_clark]

Looks like CD drive errors.
Search the forum as this as been addressed before.
If TC appears to be running correctly, they can be ignored.

[Rich]

Hi Chan
Several other people have reported similar problems, see:
http://forum.tinycorelinux.net/index.php/topic,11413.msg60369.html#msg60369
and
http://forum.tinycorelinux.net/index.php/topic,9743.msg53090.html#msg53090
They have had better results following these instructions for burning discs
http://www.troubleshooters.com/linux/coasterless.htm

[Chan]

Looks like CD drive errors.
Search the forum as this as been addressed before.
If TC appears to be running correctly, they can be ignored.

I did do a search and couldn't find anything that helped.  Tried 3 different cdroms, 2 different hard drives, and downloaded ISO 4 different times, burned 4 cd''s with the last one burned at 8x speed.  still get the same errors.  Even had my neighbour try on his computer and he got errors as shown in the jpg I attached to initial post.  This is getting frustrating.

I am new to linux and was just trying to get tinycore working on an older machine I had.

[Rich]

Hi Chan
You don't say, but does the boot process complete? If it does, install the  cdrtools.tcz  extension
and then follow the directions given here
http://www.troubleshooters.com/linux/coasterless.htm
Also, are you certain the size of the ISO file is correct? The block number in those errors suggest
that the drive is seeking out to the 100Mbyte mark on the CD.

[gwalther]

Hi Chan
Which program did you use to burn iso?
 If in windows, I  suggest you try PowerIso...it got rid of nusinance errors for me.

[maro]

Well, I understand the theory that an issue with the ISO burning process can cause reading problems during the boot phase. I'm also aware that some users have resolved their issues by tweaking that burn process. OTOH I personally have come across CD-ROM reading issues with one particular device that appeared to be independent of the ISO burning software and OS combination. In my case it was an older system and it was behaving not too well under TC 2.x, really badly with TC 3.x, but surprisingly well when booting TC 4.0rc1 (and one can read more of the "gory" details here).

As a consequence of that testing I would suggest to consider two alternative scenarios instead of trying to test out all possible combinations of CD-ROM media, CD-ROM burner, and respective software (as all this could well end up in a "wild goose chase"):

• using boot code 'ide-core.nodma' (i.e. 'ide-core.nodma=1.1' for '/dev/hdd' or other numerical values in case of a different device) when booting MC 3.x (or MC 2.x) or (as TC 3.x has just a few hours ago been officially declared to be EOL),
• using the "hot of the press" MC 4.0 ISO image as an alternative (it's so "fresh" that the downloads page has not yet caught up to the fact and the official announcement is still being prepared by the "marketing department").

I am fully aware when posting these suggestions that it is difficult to provide general advice. Too often we are guided by our own experience. And we therefore might behave like the proverbial man which only owns a hammer, and to whom each problem looks like a nail ...


EDIT: Change of the device name for the boot code example.

[Rich]

Hi maro

Too often we are guided by our own experience.

Fortunately I've never experienced this problem.
I made my original recommendation (second link in reply#2) because I did the math. Two of the
posters listed the relevant section of dmesg and the version of Tinycore being used. Multiplying
the failing block number by 2K yielded 10.6M and 10.3M for TC3.6 and TC3.8.2 respectively.
Combined with the fact that it was only failing on one block suggested it was trying to read beyond
the last written block.
With your drive, dmesg flagged DMA errors, and in the 3.8.4 case, listed some sector and block
numbers that are totally out of wack, and probably DMA induced.
In Chan's case, there are no DMA errors being reported, just seek errors, and always on the same
block. Why it is attempting to read that far into the disk, I don't know. Regardless, reading beyond the
last block appears to be a known issue, and starting with a procedure that will burn a clean disk will
eliminate one variable to the problem.
By the way, I own several hammers and have never attempted to use them to solve a CD problem, yet.

[maro]

Rich, good points. I have to admit that I did not read the screen dumps as carefully as I should have done. But luckily you were not that sloppy.

Maybe to have a bit more of a "controlled experiment" here is something else that could be tried: I'm pretty sure that the issue occurs when the boot process checks all storage devices (incl. the CD-ROM in question) for a 'tce' directory or a backup file. Therefore this can be avoided by suppressing that search via boot codes 'base norestore'. A poor man's alternative to entering those boot codes would be to eject the CD-ROM as soon as the colourful "Booting tinycore_3.8.4" (or similar) message appears at the top of the screen. This way one should get to have at least the boot process succeed.

It might then be possible to test the CD-ROM device by running a checksum test either for the same or a different media (e.g. via md5sum /dev/hdd). It might OTOH be even better (if the RAM or other disk situation allows for it) to copy the entire CD-ROM into a file (e.g. via dd if=/dev/hdd of=hdd.img) as that would allow any oddities with identifying the end of the written image to become more obvious.

In my experience it is best to have multiple shells available when undertaking these steps as the reading process might get stuck completely. So using TC or multiple virtual terminals with MC (e.g. via sudo openvt -c 2 su tc) has been the way to go for myself in this situation.

[bmarkus]

Now as TC 3.x reached its EOL, worth to move to TC 4.0 as the official version.