Description of Problem: Trying to install 7.1 on a NEC 6260 Laptop Pentium 266 MHz, with 96 MB RAM, 20x CD ROM, 5 GB HD. Doing a custom installation with 1 GB Win95, 128 MB Swap rest / Installing everything. Installation proceeds fine until we get error message: "The file ... cannot be opened. This is due to a missing file, a bad package, or bad media. Press <return> to try again." This happens everytime with repeated attempts using two different new boxed set CD Roms (one of which has already been used to do a sucessful installation on a different laptop). Failure occurs at different places between 70 MB and 800 MB into the installation. The laptop had been running RedHat 5.1 successfully for 3 years (VA research: Varbook 90 installation). Error messages from consol: <7> ISO966 Extensions: RRIP_1991A <4> Unable to identify CD-ROM format . (10 seemingly irrelevant lines) . <4> hdb:command error: status=0x51 {DriveReady SeekComplete Error} <4> hdb:command error: error=0x51 <4> end_request: I/O error, dev 03:40 (hdb), sector 272 Spoke with support - they suggested a bugzilla report. How Reproducible: Absolutely Perfectly Steps to Reproduce: 1. Boot Redhat 7.1 installation CD 2. select custom install, 3. begin installation - wait Actual Results: Fails as above Expected Results: Successful installation: joy and happiness follow Additional Information: Help please!
Can you try to boot the installer with 'linux ide=nodma'? Does that help at all?
I finally managed an installation before getting bfox's suggestion. I eliminated all but the packages I know I'll need to get the installation size down to 880MB and tried to install that. It failed several times as expected but at different points in the installation. Eventually it managed to run long enough to get the 880MB installed so at least I have a working system now. I don't want to wipe this out to try again. I do not think the bug is fixed.
I think that this is another occurence of the DMA bug that has affected a bunch of people. I think that if you had done the install with 'linux ide=nodma' then you would have avoided the problem. Basically, here's the deal. The kernel tries to use DMA transfers if the devices can support it. Most drives that can't support DMA fall back to a lower speed. Some devices, however, don't fall back properly...they just work very strangely (or not at all). The kernel team is aware of the problem and has assembled a list of devices not to use DMA transfers on. Thanks for your bug report.