Help - Search - Members - Calendar
Full Version: Persistent Freezing In Redhat 8.0 Installation
Linuxhelp > Support > Technical Support
johnk
Hi,

I've tried install Red Hat 8.0 about 5 times now, and every time when it gets to installing the packages it freezes. The partitioning and formatting part of the install seem to work fine. It starts off copying the packages just fine, but always freezes in a different place. When it freezes it doesn't attempt to access the cd drive, and both the activity and power light stay on. The only way to get out of this state is via the power button. I tried cleaning the cd and leaving the computer off for hours so it could cool down but neither of these have helped. I've given more than adequate space to the swap, root, and boot partitions prior to this. I'm attempting to install Red Hat to a fresh 120 GB Western Digital WD1200JB hard drive (slave) and I'm using Windows 98 on a smaller (12 GB) hard drive (master). If anybody has any idea on how to help me I would appreciate it as installing Red Hat was one of my primary reasons for purchasing the new drive. If any hardware information please tell me what you need and how I can obtain it. Also, I've tried the media check and it passed.

Thanks,

John
hughesjr
You may have a problem with a 120gb hard drive and the computer's bios.

When you boot the computer, does it recognize the drive as 120gb on the boot screen?

You may need to update the BIOS for the motherboard. What motherboard do you have?

----------------
Could also be a CDROM drive hardware issue or a RAM issue.

Try memtest86 on the machine.
johnk
I think that sounds possible. At boot, it shows the hard drive as being only about 48 gigs and in fdisk (in win98) it shows it the same. The Red Hat install program recognizes the drive for the full 120 gigs though. Do you think this could still be a problem? Without sounding like too much of a beginner, where can I find my motherboard information in win98, I'm looking in the device manager and I can't seem to find it there.

John
johnk
Just ran memtest86, it passed 4 times, and it had 0 errors.
hughesjr
I think the motherboard BIOS is the problem ... but in order to update it, you need to know the manufacturer and model number ... that info will probably not be anywhere in windows.

If you have a brand name PC (dell inspiration 8100, hp pavillion 3840, etc.) then we can probably find the info from that .. if you have a generic PC then you can look on the board itself for a model .. or on bootup, write down all the info at the bottom of the screen about the BIOS version and we might be able to find it that way...
johnk
I can't seem to find the exact model number, but it is a gateway and it says 500 on the case. p3 500mhz, 12.6 GB hd, (it came with less than 160 MB ram beacuse I upgraded but I don't recall exactly how much it started with). I bought the computer maybe 5 or so years ago.

I opened it up and the only writing I found on the motherboard was:

Creative E51373
1373-0001-01

SMSC FDC37M707
B9847-D75110
2SMT39113.1

Upon booting up I got this information:

PhoenixBIOS 4.0 Release 6.0

...then some copywrite stuff followed by
4W4sB0X0.1SA.0012.P07

Also, I found out that I have a PROMISE Ultra 66 Controller I guess it is, and the disk that came with the computer has driver version 1.42, and I don't think I updated that.

Not sure if any of that information is helpful but if not please reply and I'll try to keep searching to find out more info on this motherboard.

Thanks a bunch
John
hughesjr
Motherboard Controller
Is it possible that this is the correct BIOS info for the motherboard:
4W4SB0X0.15A.0012.p07

instead of this:
4W4sB0X0.1SA.0012.P07

If so, then the BIOS can be upgraded from HERE to version 4W4SB0X0.15A.0019.P14. HERE are installation instructions for the BIOS upgrade.

--------------
Here is a Forum thread (on Tech Support Guy forums) about a different model gateway (that uses the same BIOS) ... it says drives up to 127mb work after the BIOS is upgraded...
__________________________________________________________________________

Promise Controller
Is the promise controller a seperate card? If so, when it boots, it should give you a name and bios information as well....

HERE is the latest BIOS for the Ultra 66 (if it is that model ... and a stand alone card). There are execellent instructions in the readme.rtf file included with the BIOS file (use the instructions for win98). HERE is the latest Windows driver. If you upgrade the BIOS, you should also upgrade your Win98 driver...

This BIOS is V2.00 b18 ... when you boot you should see the current bios for the Ultra 66 (probably after the initial startup screen and motherboard bios info and before win98 loads).

Your hard drive performance (for the new drive) should be better with the drive connected to the promise controller (on a seperate channel if possible)...however, booting could be tricky that way.
johnk
You were right about the S being a 5. When I tried to update the BIOS on my motherboard, everything went well until I restarted. It gave me 2 checksum errors, one on my system password so it cleared it and the other on my CMOS setup. I thought that was odd so I went into setup and didn't find anything fishy. Everything booted normally from there and when I restarted I didn't find the errors again. It now shows the updated BIOS version when I boot my computer and everything runs as normal.

I was successfully able to update my Ultra 66 BIOS/Drivers.

After updating my Ultra66 BIOS/Drivers, on boot it shows the drives full capacity at boot. I'm not sure if it matters, but when I use fdisk in win98 it still shows the drive as being about 48 GB. When I go into my CMOS setup, under IDE configuration under Primary IDE Master it has none and then when I press enter on this it has [auto] on. It also has the same settings for Primary IDE slave. It was this way before and after both BIOS upgrades.

The bad news is that after another attempt to install Red Hat, the same freezing occured during the same part of the install. Since I'm getting desperate (and I don't want to resort to installing a new CD drive), I should mention that when I boot the computer, there is the first POST beep, followed by a pause, followed by 4 consecutive POST beeps. I went to the Phoenix website and tried to figure out what is wrong but I couldn't find out what this meant because the .pdf chart was too complicated for me to understand. I'm thinking there could be hardware trouble and this could possibly be giving me my problems.

If you have any more suggestions for me I'd be glad to give them a try. If I wasn't clear or you need more information I'd be glad to give it to you.

Thanks

John
johnk
Update: I just fixed my POST beeping problem. Turns out I had a bunch of garbage symbols in my Autoexec.bat file. Once I removed them, it only does the normal single test beep.
hughesjr
very strange ... that normally means a hard drive write error, a cdrom error, or a read error.

How are you doing the partitioning ....

try a 100meg boot partition as hdx1, a 1gb swap partition as hdx2 and and whatever you want as a / partition (at least 8gb ... but maybe more, you can use all of it if you want).
johnk
I tried and it froze again. This time I tried conrol-alt-f1 and the screen was filled with this message (which seems like it could be a problem):

(anaconda : 81) : GnomeCanvas-CRITICAL **:file gnome-canvas-path-def .c:line 1142
(gnome_canvas_path_def_any_closed):assertion 'path != NULL' failed

Is it possible there's a bug in the installation program? That looks like some sort of programming error to me. I don't know how much that means, but do you think that rules out the bad CD drive possiblity?
johnk
According to this post:

http://www.unixquestions.org/questions/history/32202

Somebody gets an error similar to mine, and somebody responds that it is bug 61917. I guess trying to install a newer version of Red Hat would be a good idea.
hughesjr
The RedHat 8 support EOL is 12/31/2003 ... so RedHat isn't obligated to provide any security updates after tommorrow anyway.

RedHat 9.0 ends it's support in April 2004. From the redhat website:

QUOTE (http://www.redhat.com/apps/support/errata/)
Red Hat Linux -- Red Hat's policy for Red Hat Linux distributions is to provide maintenance for at least 12 months. At certain times, Red Hat may extend errata maintenance for certain popular releases of the operating system. End of Life dates for errata maintenance for currently supported products are listed below:

  Red Hat Linux 9 (Shrike)    April 30, 2004
  Red Hat Linux 8.0 (Psyche)    December 31, 2003
  Red Hat Linux 7.3 (Valhalla)    December 31, 2003
  Red Hat Linux 7.2 (Enigma)    December 31, 2003
  Red Hat Linux 7.1 (Seawolf)    December 31, 2003

Because of that, I would not recommend RedHat Linux 9.0 for anything other than learning...

RedHat would have you either go to FedoraCore or RedHat Enterprise Linux {see this migration help article...and This Page}.

Fedora has issues (I don't think it is ready for real deployments ... and Joey (the maintainer of this site) installed, then removed Fedora from his machine. I have a Fedora drive setup for one of my test machines, but I wouldn't really recommened it. Too many small issues (in my opinion).

That narrows down the RedHat choices to the RHEL ... which is stable and a good product, but costs money (every year) to get updates ... and requires you to accept a contract to pay for a subscription for each box the product is installed on (and to renew it yearly) ... good if you need the support for every box, but certainly costly.

-----------BUT there is another solution that is based on the RHEL source code------------

There is a project called WhiteBox Enterprise Linux that is based on SRPMS from RHEL 3. They have a good track record for releasing ERRATA for WhiteBox within 24 hours of it's release for RHEL 3 from redhat.

I am running this OS on 5 boxes and I am running RHEL 3 ES on one box ... I can't tell the difference except with a uname -r. My supported RHEL box is Oracle database server ... I run a backup of that database on one of the WhiteBox boxes (an export goes into the backup server everyday). The boxes are the same hardware and the databases perform identically....

Here is a thread on freely distributable Enterprise Linux solutions. It also covers another RHEL based distro called CentOS-3 ... which is not released as a final product yet, but also has great potential.

As time passes, products based on RHEL 3 will be stable ... but they will become outdated. If you want to test all the bells and whistles, and are willing to sacrifice some stability for that, then Fedora Core is the right RedHat choice.

All in all, I would say .... give Whitebox a spin!

Here are some whitebox articles:

http://www.osnews.com/comment.php?news_id=5443

http://linuxtoday.com/developer/2003121902626NWRL

http://www.distrowatch.com/?newsid=01199

http://lwn.net/Articles/63478/

{to be fair, chrisw (another one of the admins for this site), is having problems with the network card startup of a tulip based network card on WhiteBox and had to add a script to restart his network card ... but I haven't had any real problems on my machines}
johnk
Update: For now I'm glad to annoucne that I've successfully installed and am running Red Hat 9.0! Thanks for your help and advice, I might look into other distrobutions after i get a little practice with this.

However, I noticed it was a little slow booting up and it keeps repeating these messages:

hdf: dma_timer_expiry: dma status == 0x61

PDC202XX:Primary Channel Reset

hdf: timeout waiting for DMA

hdf is the drive with red hat linux on it. I didn't set up anything for DMA, it must have done so automatically. Stuff loads a little slow and it all seems related. Just wondering if there was anything I could do about that, this new drive is supposed to rival high-end SCSI drives in speed! I did the redhat update, but that didn't do much for this problem.

John
hughesjr
post the output from these commands:

/sbin/hdparm -i /dev/hdf

and

/sbin/hdparm -v /dev/hdf
johnk
/sbin/hdparm -i /dev/hdf

/dev/hdf:

Model=WDC WD1200JB-00FUA0, FwRev=15.05R15, SerialNo=WD-WCAES1016450
Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=74
BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=16
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=234441648
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5
AdvancedPM=no WriteCache=enabled
Drive conforms to: device does not report version: 1 2 3 4 5 6

/sbin/hdparm -v /dev/hdf

/dev/hdf:
multcount = 16 (on)
IO_support = 0 (default 16-bit)
unmaskirq = 0 (off)
using_dma = 0 (off)
keepsettings = 0 (off)
readonly = 0 (off)
readahead = 8 (on)
geometry = 14593/255/63, sectors = 234441648, start = 0


Here is the exact drive here if this helps also: http://www.zipzoomfly.com/jsp/ProductDetai...e=101205-315002
(Western Digital 1200JB)

John
hughesjr
You have no DMA and the controller is constantly resetting ... not good
------------------
I have found some info on the pdc202xx driver and redhat 9, here is a listed bug that seems similar ... but they don't say they fixed it.

I would recommend that you download and install the newest redhat9 kernel:

kernel-2.4.20-27.9.i686.rpm

and install it like this:

(from within the directory you download to as the root user):

rpm -Uvh kernel-2.4.20-27.9.i686.rpm

reboot and select the new kernel and see if the controller works. If you have also installed kernel-source (do the command rpm -q kernel-source )... if it returns a value then download 2.4.20-27.9.i386.rpm and install like this:

rpm -Uvh 2.4.20-27.9.i386.rpm
johnk
I believe I already tried to update the kernel -
I ran the Red Hat Update Agent and updated everything except the kernel. When I run the agent, it lists the kernel under 'Packages Flagged To Be Skipped'. Under reason skipped it has Pkg name/pattern. Anyhow, I figured updating the kernel was an important step so I selected it, and continued with the update/install. After a reboot, I wasn't able to boot into either version of Red Hat because when it got to a certain step (I beleive it was checking filesystem) it said it had to run a check and then it said FAILED. The agent lists the kernel as version 2.4.20 and the release as 27.9. So, since I wasn't able to boot into either version and lack the skills to be tinkering around, I just decided to do a fresh install and here I am. It sounds like what you're saying is a differnet way to do the same thing, right? If so, I might run into a similar problem and be stuck doing another fresh install. I could try the update and then boot into windows and post back and forth if you think we could get it up and running, but I don't think you'd want to waste all that time trying to help me with what seems like a real muddy problem. =) I still have a lot of room on this drive, and I thought about trying a differnet distribution that might not have the same bug. Do you think it would be a good idea to try another to see if I have a similar problem? What do you think would be my best bet?

John
hughesjr
I have seen the dma reset problems many times when a hard drive fails .... but not with a controller card issue.

Are you using one of the newer IDE cables with the drive (80 wires instead of 40) ... here are the two cable types:

40 Wire, 40 pin (older, < ATA66) IDE cable

--------------------------------------------------------------------
80 Wire, 40 pin (newer, for ATA66 and higher) IDE cable
[img]http://www.hughesjr.com/stuff/idecable80.jpg[/img]

Also, for the 80 wire cable, make sure to use the Grey (middle) connector for slave and the black (end) connector for master drives...the blue connector is for the controller.

I would try connecting the drive to one of the motherboard controllers (then re-installing Linux) and see if you get the same problems.

When there is a hard drive failure, this is also how it looks in Linux. If it exhibits the same behavior on the Motherboard controller (after a fresh install) then this is not a Ultra66 issue, but a hard drive issue ...

If it is a hard drive issue, I would use fdisk to remove all the partitions and take it back to the place you got it for a replacement (if it is fairly new, OR contact the manufacturer and ask for a replacement if it is under warranty.
johnk
I moved the drive off of the ultra 66 controller and onto the motherboard controller (as master). This solved all issues, and I've decided to go with slackware 9.1 if you were curious. Thanks a lot for your help! But, since I'm here I did have a few questions about configuration... =)

1. During the slackware install I tried to install LILO and it said LILO returned an error and it wasn't successful. I've tried booting from a boot disk and running liloconfig, but whether I try to install to MBR or linux partition I get the same error. (I'd prefer not to have to boot from a boot disk every time). The root partition is flagged as bootable in cfdisk. GRUB used to be my boot loader from the red hat installation, but I've done a fdisk /mbr since then and I believe it removed that (and it doesn't show up anymore on boot).

2. I skipped the network configuration option during the install because I had been there in one of my previous installs and didn't know what to put for certain values. I'm trying to set up my ethernet card, and I can't find the option in GNOME. I'd prefer to set it up in GNOME as opposed to the command line, how can I do this?

Thanks again!
John
johnk
Error #1 to be exact.
hughesjr
Try booting into linux and running the command:

/sbin/lilo -v -v

to rewrite LILO to the MBR.

(make sure in the computer's BIOS you don't have any settings enabled that lock the MBR to prevent it from being written to .... a Virus Protection selection in the bios is one thing that does this).

I'm installing slackware 9.1 on a partition right now (I haven't used LILO in a while so I want to see it in action).
johnk
I just figured out how to tinker around with LILO and got it working a little, but I seem to prefer GRUB much more. I looked in my /sbin folder and don't see anything grub related there (I don't think slackware gave me the option). Could you explain to me how to install GRUB instead? I'm sure it must be somewhere on that Red hat install cd....
hughesjr
I don't like slackware all that much, mainly because I don't like it's package management system or it's runlevel startup system.

For package updates, I highly recommend swaret if you are going to use slack.

I am going to finish up my test slackware 9.1 install in a couple minutes and install SWareT 1.6.1_test2. I'll then install grub and post how I did it....
hughesjr
OK,

I finished my Slackware 9.1 installed and installed SWareT 1.6.1_test2. How to install SWareT:

1. Download SWareT from HERE.

2. Use pkgtool (type pkgtool a terminal window {as root} from within the directory where you downloaded swaret) ... then select and install swaret.

3. Configure /etc/swaret.conf ... my current configuration is here. My settings are fairly conservative ... I have VERSION=9.1 instead of VERSION=current ... Also all the updates will only be done from official Slackware mirrors (that are fairly fast to the USA) ...

I added (but remarked out) some Slackware Repositories for packages other than those provided officially by Slackware. (I normally only unremark one of these repositories when I want to get a package ... then immediately remark it back out when I'm done. We will do this later to install grub!)

4. After /etc/swaret.conf is configured, run the command:

swaret --update

to get information and discreptions for the latest slackware 9.1 packages.

5. Issue the command:

swaret --upgrade

to upgrade all installed packages. The packages that are excluded from /etc/swaret.conf (like kernel, lilo, swaret) will not be updated. I always make sure not to have non-slackware repositories present inf /etc/swaret.conf when I do swaret --upgrade.

6. Restart Slackware after all the upgrades are finished to make sure all is well....

------------------------------
Installing grub

1. Edit /etc/swaret.conf and enable the linuxpackages.net add-on repository (by removing the #) to make the line look like:

REPOS_ROOT=LinuxPackages%http://www3.linuxpackages.net/Slackware-9.1

2. Update the swaret packages with the command:

swaret --update

3. Install the latest grub package for slackware with the command:

swaret --install grub

4. After the package is installed, remark out (by placing a # in front of) the line we unremarked earlier:

#REPOS_ROOT=LinuxPackages%http://www3.linuxpackages.net/Slackware-9.1

5. Update the swaret packages (with linuxpackages.net removed) with this command:

swaret --update

Important: If you don't do steps 4 and 5, you will get non-official slackware packages (that can over write the official slack packages) from linuxpackages.net.

6. Install grub with the command:

grub-install /dev/hda

(substitute the correct drive if not installing on the primary master)

7. Edit the file /boot/grub/menu.lst and set it up correctly....

Mine looks like this (I have slackware installed with no boot partition and root partition installed on /dev/hda6):

CODE
# Sample boot menu configuration file
#
                                                                               
# Boot automatically after a minute.
timeout 5
                                                                               
# By default, boot the second entry.
default 0
                                                                               
# Fallback to the first entry.
#fallback 0
                                                                               
# For booting Linux
title Slackware9.1
 root (hd0,5)
 kernel /boot/vmlinuz root=/dev/hda6 vga=773


I'm not sure what your partitions are ... but if you had a boot partition as hda1, swap as hda2, and root as hda3 ... menu.lst would look like this:

CODE
# Sample boot menu configuration file
#
                                                                               
# Boot automatically after a minute.
timeout 5
                                                                               
# By default, boot the second entry.
default 0
                                                                               
# Fallback to the first entry.
#fallback 0
                                                                               
# For booting Linux
title Slackware9.1
root (hd0,0)
kernel /vmlinuz root=/dev/hda3 vga=773


The vga=773 is not required and can be left off (it sets a framebuffer mode of 1024x768 at 256 colors for the console).

8. Now to write grub to the boot record, issue this command:

grub

A. Now you should be in the grub interactive menu....if your boot partition is hda1 use the command

root (hd0,0)

(i would use root (hd0,5) since I am going to setup /dev/hda6)

B. Issue the command to setup the boot record ... I think you want to setup MBR:

setup (hd0)

I want to setup a partition boot (for /dev/hda6) and not MBR, so for me the command is setup (hd0,5).

C. If those 2 commands complete without error, now exit the grub menu with the command:

quit

9. You should now be able to boot with grub.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2017 Invision Power Services, Inc.