Bug 37280
Summary: | (IDE)random data corruption that ide=nodma does not fix | ||||||
---|---|---|---|---|---|---|---|
Product: | [Retired] Red Hat Linux | Reporter: | Stan Stead <sstead> | ||||
Component: | kernel | Assignee: | Arjan van de Ven <arjanv> | ||||
Status: | CLOSED CURRENTRELEASE | QA Contact: | Brock Organ <borgan> | ||||
Severity: | high | Docs Contact: | |||||
Priority: | high | ||||||
Version: | 7.2 | CC: | cbailes, cl, crowed, david.schreiber, doug_reed, ewt, infotech, kthomas6, ldunaway, mblm96, mintvelt, rkane4, taylor, toby.maxwell-lyte | ||||
Target Milestone: | --- | ||||||
Target Release: | --- | ||||||
Hardware: | i386 | ||||||
OS: | Linux | ||||||
Whiteboard: | |||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | Environment: | ||||||
Last Closed: | 2004-09-30 15:38:57 UTC | Type: | --- | ||||
Regression: | --- | Mount Type: | --- | ||||
Documentation: | --- | CRM: | |||||
Verified Versions: | Category: | --- | |||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
Cloudforms Team: | --- | Target Upstream Version: | |||||
Embargoed: | |||||||
Attachments: |
|
Description
Stan Stead
2001-04-23 23:20:46 UTC
Is this with the same system that saw the behavior from bug #37253? *** Bug 37573 has been marked as a duplicate of this bug. *** Sme problem here. Currently running RH 7.0 upgrading to 7.1. My machie is a Comapq Deskpro Pentium Pro 200MHz, 128MB RAM, Western Digital 6.4GB hard drive, Matrox video, IBM 4x SCSI CDRom, Python SCSI 4mm DAT (DDS-DC) anmd Adaptec 1940U SCSI host adaptor. Same exact message is received after I make it through the install process in the X GUI. Error ocurs after install searches through the package list to determine what it needs to upgrade. everything appears to be running smoothely then, CRASH! WARNING: Unknown EDD version 0x1600 supported. More information. Tried text mode installation this morning as follows: 1. boot: text 2. Received: Running anaconda - please wait... 3. Chose English 4. Chose us keyboard 5. Upon choosing OK at the Welcome to Redhat message then received... 6. Installation selection and chose Upgrade Existing Installation. 8. Chose Ok and immediately received "WARNING: Unknown EDD version 0x1600 supported" 9. Installation continued and I received "Error mounting ext2 filesystem on sdb1 - Invalid argument". Same for sdc1, sdd1 and sde1 which are all part of a RAID volume. This all makes sense so far. I can only assume that the WARNING is because of the file system on sda1 which was not displayed in the mount errors. 10. Installation continued and th message "Upgrading the Red Hat Linux Installation on partition /dev/hda11" was displayed. I chose OK. 11. Displayed was "Reading package information..." 12. Installation failed. with errors as follows: python: header.c: 241: regionSwab: Assertion `i.info.type >= 0 && ie.info.type <=9' failed. install exited abnormally -- recieved signal 6 sending termination. What gives?? ewt, do you have any ideas about this? I have the exact same problem. I am using a Compaq Deskpro 2000 5200MMX and I can not do a fresh install or an upgrade of 7.1 without getting the message "Warning: Unknown EDD version 0x1600 supported" twice. Version 7.0 installs just fine. I get various errors/symptoms like the PC hanging while trying to format partitions, DMA time-out while formatting Swap partitions, unable to format "/" partition, "error transferring install image, you are probably out of disk space", and "install exited abnormally -- received signal 11" I've tried 3 different hard drives, tried manual (using DRUID) and auto partitioning, tried 32Mb, 64MB and 96MB RAM, tried a minimum install, tried a 7.1 beta CD that worked fine on another PC, turned of LBA mode in BIOS and tried upgrading the Compaq BIOS and the 7.1 install still fails. I am currently out of ideas. OK! NOW I am getting another error. During my first attempt I made at installing using text mode I created a additional 128MB swap file in /var/. I removed refernce to it in my /etc/fstab file and rebooted. I booted from floppy (boot.img) and ran text mode installation again following same steps as before, chossing "Upgrade Existing Installation". This time I got and Exeption occurred. I tried this multiple times and got the same errors. I will attache a information dump which I was able to save to floppy. Created attachment 16565 [details]
Exception Occurred - Traceback (iinermost last) anaconda line 520 in ? intf.run...
I am disspointed by the lack of motion on this bug. What's up Redhat???? What happens if you type in 'linux ide=nodma' at the boot screen? the results of running installation using linux ide=nodma. System booted with floppy (boot.img) and CD in IBM 4x SCSI CDRom. This is the Compaq Deskpro 6000 Pentium Pro 200MHz machine. Running anaconda - please wait... Probing for mouse type... Waiting for X server to start...log located in /tmp/X.log .... X server started successfully. Gdk-ERROR **: X connection to :1.0 broken (explicit kill or server shutdown). install exited abnormally sending termination signals...done disabling swap... unmounting filesystems... /mnt/runtime /mnt/source umount failed () /dev/pts /proc you my safely reboot your system I tried typing this on the boot prompt from the Red Hat installation CD ("boot:linux ide=nodma <enter>") and it made no difference. I still get the message "Unknown EDD version 0x1600 supported" twice, and my installation still fails when trying to format my partitions. From edd.c: if ((regs.eax & 0xff00) == 0x0100) { ec->version.major = 1; ec->version.minor = 0; } else if ((regs.eax & 0xff00) == 0x2000) { ec->version.major = 2; ec->version.minor = 0; } else if ((regs.eax & 0xff00) == 0x2100) { ec->version.major = 2; ec->version.minor = 1; } else if ((regs.eax & 0xff00) == 0x3000) { ec->version.major = 3; ec->version.minor = 0; } else { fprintf (stderr, "WARNING: Unknown EDD version 0x%x supported\n", regs.eax & 0xff00); } It looks like edd.c doesn't know anything about EDD version 0x1600. Matt, do you know why this is? I have duplicated this problem on another Deskpro 2100 with nearly identical hardware. I decided to do a "clean install" of the workstation. I tried using the same exact partitions and got an error message that the swap partition was too small and the system hung again. I then rebooted and did a clean install with RH 7.1 picking the partition sizes. Everything worked correctly. I suspect that the problem is related to the need for a larger swap partition and that the message regarding the EDD version may be just a red herring. I am pleased to see that others have the ability to play with other systems and have gotten RH 7.1 to install. My problem is that I have a system to UPGRADE and thus far cannot. Again, I am dissappointed with the support from Redhat on this issue. What is EDD? It is pretty obvious that the latest comments from Redhat were, shall we say, rhetorical? "It looks like edd.c doesn't know anything about EDD version 0x1600. Matt, do you know why this is?" Did we need to wait two weeks for someone to take a look at edd.c to tell us the obvious? Is this EDD warning a red herring, and my problem is unrelated? If so do I need to open separate bug repot? I would like to upgrade my system before Christmas 2001 (which, btw, is only 7 months away) ;-) First of all, this is not a support system. If you have purchased a product and are having problems installing, you should contact our installation support department. The "WARNING: Unknown EDD version 0x1600 supported" message is a red herring, it should have nothing to do with your system hanging. These random *different* failures indicate data corription, most likely caused by IDE DMA. To turn this off, boot with "linux ide=nodma" at the boot: prompt. I understand that this is not a support system, however, it is a bug report system, which Redhat has initiated and does in some way support, otherwise.. a. You would not be speaking with us here. b. You would not be receiving paychecks from RH. As one can plainly see, we have tried booting with "linux ide=nodma" and have had the same results. Do you people read what we type here? Given that we have tried your suggestions, I ask my question again... How do we fix this problem. Given your response, the EDD warning is a red herrring, thus there must be another issue here. Are all of our problems related or are they each separate issues for which we should open new bug reports for, closing this one? Where do we go from here? How do we fix this? I have tried using the suggested fix: linux ide=nodma. The system still crashes at the same place. There is no apparent difference between using default <ENTER> and the fix. Sorry guys, this is not the problem. Me too, I also have a compaq deskpro 2000. P166MMX 64MB RAM standard CD drive that came with it. Language English UK, 102 keyboard, 2 button PS/2 mouse, London timezone. I get the following error messages: - WARNING: Unknown EDD version 0x1600 supported WARNING: Unknown EDD version 0x1600 supported Gtk-WARNING **: cannot open display: :9 Gdk-ERROR **: Fatal IO error 104 (connection reset by peer) on X server :9. python: header.c:511: headerload: Assertion `rdlen == dl' failed. install exited abnormally seding termination signals...done sending kill signals...done disabling swap... /tmp/swap/hda5 unmounting filesystems... /mnt/sysimage/boot /mnt/sysimage/home /mnt/sysimage/proc /mnt/sysimage/usr /proc/bus/usb /mnt/sysimage /mnt/runtime /mnt/source umount failed () /dev/pts /proc you may safely reboot your system I recieved a support call on this issue. Thanks Nidil! I am cc'ing him onto this bug report. I told him to type: linux ide=nodma nousb And he was able to get it booted into the installer. He has an HP Pavillion 850 or 950 Mhz Athalon processor He said he'd add comments to this report. He said he would try typing JUST nousb and JUST ide=nodma to determine if only one or both of these flags are really needed. So anyone who could not get it booted w/ ide-nodma, try nousb... Johnray out Tried nousb. Same result. Hangs at the same point. Strange question: does this machine have a mouse attached ? I tried it as well. Used "linux ide=nodma nousb" and I can at least get the X Windows installer to run, however, I still get installer failure. I chose Custom installation and partitioned manually. I was able to select packages to install and the installer formatted the partitions I created then checked package dependencies, began to install and BANG! I get the exact same python error as Toby. WARNING: Unknown EDD version 0x1600 supported WARNING: Unknown EDD version 0x1600 supported python: header.c:511: headerload: Assertion `rdlen == dl' failed. install exited abnormally -- received signal 11 etc... *** Bug 40352 has been marked as a duplicate of this bug. *** I have a Compaq DP2000 5/200MMX,3C905 NIC, Seagate 4.3 Gig Medalist HDA,Mitsume CD. I can boot the system off of the CD, run the graphical installation, set up the disk partations, select package configurations, time zone etc.. when the package installation begins the system throws an error about the disk being full. The installation allows me to restart but I can never get past this point. Doing the three-finger-windows-repair routine shows the edd error on the cons. *** Bug 37151 has been marked as a duplicate of this bug. *** OK, this shouldn't help one bit, but please download: http://people.redhat.com/msw/edd-updates.img write it out to a blank floppy. Boot the installer floppy or CD with "linux updates". It will prompt you to insert this disk. This should make the "unknown edd version" message go away. *** Bug 41392 has been marked as a duplicate of this bug. *** Doug Reed... I am reporter of bug 41392 deemed to be a duplicate of this problem. I fully agree that my problem is a duplicate of this, and originally agreed that the EDD message was likely a red herring...but I am becoming less so based on the fact that the users all seem to be reporting somewhat different symptoms with this as the most obvious common denominator. I will be further convinced if indeed the user with the HP Pavillion 850 (reported in the entry by jrfuller) can confirm that he/she also gets the EDD message. Also of interest is that other users are having problems at different stages of the install. I have several Compaq machines of a similar vintage, and each has died in exactly the same place with every version of RedHat starting with Version 6.2. Version 6.1 and earlier do not exhibit this behaviour. It is also interesting that users are reporting that 7.0 works, since I experience the same problem on both 6.2 and 7.1. I have never tested 7.0. My machines ALWAYS freeze with no error messages always in the middle of the file copy, and I have indeed already tried the 'ide=nodma' switch some time ago... I have never tried the 'nousb' switch, although others seem to report that this does not solve the problem. I would also tend to doubt that USB would be the problem based on the fact that in my case and others the problems occur well after the Hardare probing stage during file system format or copy. One would not expect the installer to do anything to upset the USB subsystem while performing file I/O on IDE devices. Question to RedHat Development! What is EDD? RedHat support told me that the EDD messages could not possibly have been generated by the RedHat system, but had to be coming from my BIOS and proved that RedHat Linux was unlikely to run. ...I did not really believe that my BIOS was generating the message, since I have never seen a PC BIOS that generated text messages to running systems and allowed them to continue running...but then, this IS a Compaq BIOS, so there really isn't much that I would eliminate. I would like confirmation that the HP system is experiencing a similar problem and that it gets the EDD message. Also, by way of constructive criticism, the reason I opened problem 41392 was the telephone support told me that the message was not being generated by the RedHat system, that the message did not show up in their database and that I should open a new bug report. This problem should heve shown up on their radar since the message is indeed in the "Summary"...someone might wish to look into why support could not connect my problem to this bug report. Anyway, I am not trying to beat anyone up, but it does take support time to close and connect problems, and I did not mention the EDD message in my original bug report (which I should have done... shame on me!), but I DID mention it with support when I first reported the problem, and they told me they looked in the system for this message and got no "hits". I will Download the EDD fix, and re-test when I get a chance and report back. From looking at "edd.c" above, I would not be surprised to see some ill effects from this message since the else case fails to initialize the major and minor versions at all. I would expect the code to initialize them with some values that are known to mean "unknown" as opposed to leaving them unitialized. EDD is a enhanced disk interface that allows access to data past the 1024 cyl limit from BIOS calls. LILO uses this to boot systems with kernels past 1024 cyl. Not all BIOSes can handle this, so we do a check in the installer. You would be surprised to see how much the usb subsystem can mess things up. The same goes for DMA access. We've never come across these machines that show the problem. With physical access to one I'm sure we could fix this problem quickly... Doug Reed...I would be most happy to give you access to one, but I am in California. Wanna visit LA? I have one six blocks from the Pacific ocean in Redondo Beach. ...Bring naked women...the ones here are all on BayWatch. Doug Reed...I have finally found time to perform the test. Indeed "msw" was correct on both counts... The update disk DID make the EDD message go away, and it did not help one bit. I did NOT do a clean install this time since this trashes my system... I said upgrade, and it locked up while the little "determining which packages to upgrade" dialog was displayed. The good news is that, since it hasn't DONE anything, I can reboot 6.1, fsck, and run the install again at will. It locks up after a good five minutes with the disk light flashing happily. I did this with the <CTRL>-<ALT>-<F1>, <CTRL>-<ALT>- <F3>, and <CTRL>-<ALT>-<F4> screens displayed. I copied what was on the screens below for all to see. I hope I get it all copied correctly since I have only one monitor connected to an OmniCube, I have to switch over, read a line, then switch here, and type it (Since the system is hung, and has no printer installed anyway, I cannot capture it). It does seem like IRQ 11 is a bit more popular than it should be to me in the F4 screen. ...anyway, hope it helps... I REALLY NEED THIS TO WORK! <CTRL>-<ALT>-<F1> contain: Running anaconda Probing for mouse type Starting the X server ... ...X-Server started successfully. <CTRL>-<ALT>-<F3> ends with: * probing for ide floppies * no ide floppies found * looking for video cards requiring agpgart module * found video card controller unknown * in start Pcmcia() * pcmcia probe returned: |PCI bridge probe not found Intel PCIC probe: not found Databook TCIC-2 probe: not found | * no pcic contoller found * probing busses * finished bus probing * found suggestion of usb-uhci * found nothing * trying to mount device hdc * loopfd is 7 * getting ready to spawn shell now * probing busses * finished bus probing * found suggestion of tlan * found tlan device * found suggestion of usb-uhci * found device justProbe is 0 * going to insmod tlan.o (path is NULL) * UPDATES floppy device is fd0 * recursively copying /tmp/update-disk/lost+found * going to insmod raid0.o (path is NULL) * going to insmod raid1.o (path is NULL) * going to insmod xor.o (path is NULL) * going to insmod raid5.o (path is NULL) * going to insmod fat.o (path is NULL) * going to insmod msdos.o (path is NULL) * going to insmod vfat.o (path is NULL) * going to insmod ext3.o (path is NULL) * looking for USB mouse * no IDE floppy devices found. <CTRL>-<ALT>-<F4> ends with: <6>IP Protocols ICMB UDP TCP <4>IP routing cache hash table of 1024 buckets, 8Kbytes <4>TCP Hash table configured (Established 8192 bind 8192) <6>NET4 Unix domain sockets 1.0/SMP for Linux NET4.0 <4>VFS mouted root (ext2 filesystem) <4>hdc ATAPI 20X CD-ROM deiver, 128Kb cache, DMA <6>Uniform CD-ROM Revision 3.12 <6>SCSI subsystem driver revision 1.00 <6>usb.c Registered new driver usbdevfs <6>usb.c Registered new driver hub <6>usb-uhci.c $Revision 1.251 $ Time 18:28:33 Apr 8, 2001 <6>usb-uhci.c High bandwidth mode enabled <6>PCI Found IRQ 11 for device 00.14.2 <6>PCI The same IRQ used for device 00.0d.0 <6>usb-uhci.c USB UHCI at I/O 0x1000, IRQ 11 <6>usb-uhci.c Detected 2 ports <6>usb.c New USB bus registered, assigned bus number 1 <6>hub.c USB HUB found <6>hub.c 2 ports detected <6>usb.c Registered new driver hid <7>ISO 9660 Extensions: RRIP_1991A <4>Unable to identify CD-ROM format <6>ThunderLAN driver v1.14a <6>PCI Found IRQ 11 for device 00.10.0 <6>TLAN eth0 irq=11, io=1020, Compaq Netelligent 10/100 TX Embedded UTP, Rev. 16 <6>TLAN 1 device installed, PCI: 1 EISA: 0 <6>raid0 personality registered as nr 2 <6>raid1 personality registered as nr 3 <6>raid5 measuring checksum speed <4> 8 regs : 251.600 MB/sec <4> 32 regs : 208.000 MB/sec <4> pII_mmx : 344.800 MB/sec <4> p5_mmx : 420.400 MB/sec <4>raid5 using function: p5_mmx (420.400 MB/sec) <6>raid5 personality registered as nr 4 <6>Adding Swap 266072K swap-space (Priority -1) I have to echo Doug Reeds comments. I too, took the time to download the update, make a floppy and go through the update process again. Like Doug, the updates eliminated the EDD message and do NOT fix the hanging problem. I too, hang at the point of customizing the upgrade list of files. I can also cycle through various screens Ctrl-Alt-F1..F10, with very similar output. No files are written, a fsck restores the system to running RH 7.0. I have also tried: linux updates ide=nodma nousb, instead of linux updates alone. Same results. Yes, I have a compaq ps/2 mouse installed. Stan I reported bug 40352 which was labels as a dupe of this. (Same problem) If you have some one in Dallas, Tx they are welcome to stop by.;) Mine is a generic duel processor MB. No Compaq, or HP. "OEM ID: INTEL Product ID: 440LX APIC at: 0xFEE00000" Doug Reed...Infonet. I dug up a copy of Mandrake 8.0 as a test, and confirmed that my test machine hangs during format in this release. This may eliminate Anaconda as a cause. Which Hard disk controller does everyone have? My Compaq's have built-in VIA IDE controllers, and I have found some fixes in the Kernel news for 2.4.5-ac6 "Much updated VIA IDE support" with regard to this controller. I notice that the original reporter has a VIA controller. Others report that 7.0 worked. Were there VIA controller fixes in the 2.2 kernel between 6.2 and 7.0 that may have fallen off, or designed out of existence in the 2.4 Kernel? If so, It may be useful to try to install using the newer Kernel with the VIA updates. I would have no idea how to do this...I would guess that someone at RedHat would have to build me(us?) a CD with the newer Kernel. I think Doug Reed might beright. My Compaq's also have built-in VIA IDE controllers. Is there anyway to us an external IDE controller on these compaq boards ? You can add me to the list as well. Both the systems I am having problems with have VIA IDE Controllers. Although sometimes I get some output stating ide0 reset error 0x080 after it locks. Both the systems I am having the problem on also have maxtor drives as hda(one 15gig and one 30gig) Has anyone, and that includes you RedHat, found anyway around this problem yet. If I go out and buy an official copy of 7.1 will RedHat resolve this problem for me ? I appear to be running into the same problem, however I usually make it past the formatting, just prior to copying the disk image or during the installation of packages the install fails and I receive a "python: header.c:511: headerload: Assertion `rdlen == dl' failed" error or it simply hangs. I do not receive a GDK or EDD error. Under Ctrl-Alt-F4 I see several instances of: "<6>cdrom: open failed <4>hdc: packaet command error=0x51 {DriveReady SeekComplete Error} <4>hdc: packaet command error=0x54" and one instance of: "hdb: lost interrupt hdb: cdrom_read_intr: data underrun (4 blocks)" The last line is: "adding swap-space (priority -1)" Underrun is a buffering error, isn't it? I have tried ide=nodma. the rtld errors are classic signs of data corruption. If ide=nodma didn't help, I'm out of ideas. I'm going to punt to the kernel team. Smells like a hardware issue.. Are the cables properly connected ? I successfully installed, it looks like a memory dump. In my bios agp is assigned 64M of ram, I have 128M so I assigned the kernel the remaining 64M with "mem=64M". I also specified "ide=nodma" but I don't think it was needed. I was receiving the "python: header.c:511: headerLoad: Assertion 'rdlen == dl' failed" error in addition to the occational signal 11, 6 or 7 error and "error loading shared libraries [...] GLIB_2.0 no defined". Is there a fix/workaround for 37280 yet ? I have hit exactly the same problem . A Compaq Deskpro that run RH7.0 with no problems fails during instalation of RH7.1. I won't repeat the symptoms, as they appear to be exactly the same. The CD set was purchased from my local chain. It was called the RedHat 'Garage' set as only contained the CD's. I was up until 4am this morning working trying to get this to install. you can always try to "dd if=/dev/cdrom of=disc1.iso" (same for disk2) and do a harddisk install. If that works, your cdrom drive REALLY dislikes dma. I have seen a similar problem with Red Hat 7.1 on the Compax Deskpro 5000. After correctly formatting my /boot partition the configuration appeared to stop operating whilst formatting the 303MB /home partition/ However my system messages (ctrl-alt-F4) point to the system continually waiting for the harddisk. It's almost as if it has gone to sleep. N.B. I have disabled power management (disabling ACPI will stop your machine booting) <4>hda: status timeout: status=0x80 {busy} <4>hda: drive not ready for command <4>ide0: reset timed-out status=0x80 <4>end_request: I/O error, dev 03:06 (hda) sector 524392 This gradually counts up at the rate of approx 2 sectors per 30 seconds. I believe that the system had already correctly detected that no dma was required. I did not download Roswell, because no such thing or place exists, and I did not try to install it on the @#* Compaq DP4000 with the @#* VIA IDE controller because no company would ever build such a system. But If HAD, IT would have locked up earlier this time... When one selects "Manually partition using Disk Druid", and clicks on "Next >"... At THIS point, the "Next >" button would ooze this tiny bit of glue, which would cause the little magic arrow to get stuck on the button, and no amount of mouse movement would be able to break it loose. Only the mysterious button on the front of the case that causes the little magic arrow to disappear completely seemed to be able to disolve this powerful glue. Some of us wish there WAS a Roswell because of all of the VIA Controller fixes that appear to be occurring in the Real World... or at least the Alan Cox world. Some of us were hopeing that Roswell would someday exist and contain such fixes and that a mysterious man with a Red Hat would see that all was well in the Compaq World. ...alas if only my company had listened to me and run the other way and hid under the bed when the evil word 'Compaq' was uttered in the presence of mortals...only then would the world be free of the evil spell, and the mysterious man in the Red Hat would no longer be responsible only for keeping coffee stains from attacking my desk and leaving the Evil Billy Borg free to roam the corridors of my office building consuming all intellectual property. Add me to the list as well. A Compaq Deskpro 166 that runs RH6.2 with no problems fails during instalation of RH7.1. Also with the VIA chipset. WARNING: Unknown EDD version 0x1600 supported etc... I've tried all permutations of ide=nodma and nousb and sometimes the system will install but will hang at a random point in the future, normally when accessing the IDE CDRom. Interestingly the disable USB in the hardware section of Compaq setup has little effect but the disable USB in the security section of the setup does make it disappear completely e.g. no IRQ usage, not detected by lspci etc. However, this does not fix the problems with the boot. Hardware Experiment: I tried replacing all the two IDE cables with UDMA 100 cables and no success. However, I did try making the CDROM as a slave drive off the primary IDE controller with the origional IDE cable. I have done a couple of installs with no kernel settlngs and no problems. The install correctly detects the hard disk is running as UDMA33 an the CDRom is running PIO. As an aside, this also appears to be the only hardware configuration which allows a boot from CDROM. The Kernel with my Redhat 7.1 appears to be 2.4.2-2. There are changes/enhancements for the VIA chipset used by these Compaq IDE controllers in Kernel versions 2.4.4 and 2.4.10. Anyone care to point me at a Redhat 7.1 boot image using a later kernel, I'll be happy to give it a go. I can confirm that making the CDROM as a slave drive off the primary IDE controller appears to work. Greetings: Check out bug 55592 and my last post. I've had similar problems. To that end could somebody email how to set my CDROM as the primary slave drive off of the IDE? I'll give that a try, but not sure how! Thanks, Kevin Has anybody had any luck with 7.2? Does it fix the problem? We don't have a Deskpro 2000 in house, so I can't test it myself. *** Bug 53547 has been marked as a duplicate of this bug. *** I am the reporter of bug 52547. I ran in to this problem when I replaced the standard 1Gb western digital hd with a matrox 20 GB. I can do a clean install on my deskpro 2000/166mmx but only if I choose not to use LILO and to create a bootflop. Funny thing. When I tried to install with both LILO and Bootdisk option turned off the install failed. When the machine runs (using the bootflop) it rains CRC errors from the HD from time to time. Hope this info helps Greetings: I would be more than happy to get a copy of 7.2 and test it on my machine to see if it works. Kevin Greetings: I just reported in bug 55592 that making the CDROM a slave of the IDE controller as reported in this bug was a success in installing 7.1 server. Just to test further, I wiped the drive and installed workstation with no problems either. I'm still willing to get a copy of 7.2 and test to make sure the kernal is working properly. Thanks, Kevin This is in response to "BFox"s request for information on Version 7.2. I have purchased 7.2 and tried to install it. There is no change. The system hangs in the same way at the same place 100% of the time in my testing without regard to which, if any, kernel switches are used. I gave up and bought a new motherboard re-using all other components. 7.2 installs with no problem. Sadly I still have several boat anchors in my data center that I would like to run something newer than 6.1 on, and cannot do so. In reply to comment made by "mintvelt". I too can confirm that the problem is in some way related to either the replacement, or the addition of an additional hard disk. Indeed the machines come with a 1 GIG drive. Some of my machines have had it replaced, and some have had either a second IDE, or a SCSI card with another hard drive installed. In every case, the new drive is larger than the 1 GIG that came with the machine, and this seems to be related to the problem, as it has problems during the file copy stage of the install. I am not sure it has anything to do with LILO, however, because 7.2 no longer defaults to LILO, and I took this default on my install, and also because 6.1 DOES use LILO, and it installs without problems 100% of the time. (I have done SEVERAL re-installations while testing this problem) *** Bug 55592 has been marked as a duplicate of this bug. *** The fact is that the Compaq Deskpro 2000 has no true bios. The bios is built into the original hard drive as a 4mb partition, which controls all onboard systems. If this partition is removed it is almost impossible to recover. HP/Compaq has rompaq which can update up to an 8.4 gb drive with the proper partition for this, however it is next to impossible to create using their system. I have been able to duplicate this with many flavors of linux and other OS versions. I hope this helps in some way. you can view some of the documentation here Rompaqs and divers: http://h18007.www1.hp.com/support/files/Deskpro/us/locate/3_65.html Documentation: http://www.brlug.net/pipermail/general_brlug.net/2000- December/000164.html Thanks for the bug report. However, Red Hat no longer maintains this version of the product. Please upgrade to the latest version and open a new bug if the problem persists. The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, and if you believe this bug is interesting to them, please report the problem in the bug tracker at: http://bugzilla.fedora.us/ |