From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 4.0) All Instances are booting from 7.1 cdrom (floppy behaves the same) no kernel params - fails; insufficient memory. linux mem=265M - fails; cannot mount root fs (report errors about loading init.img past end of ram... only appears to detect 16meg) linux mem=exactmap mem=256M@0 - fails; mke2fs crashes system with the following trace... (crashes both with auto partition/format and manual partition/format) Hardware is assumed good as 7.0 installs correctly without errors and functions normally. Hardware Descrption: Compaq Proliant 2500 BIOS: E24 - 12/01/1999 (most current on compaq web site) RAM: 256Meg CPU: Dual Pentium 200 Compaq Smart Array 2dh controller 3 4.3 gig drives in mirror w/spare config Reproducible: Always Steps to Reproduce: 1. Boot redhat 7.1 cdrom 2. at boot: prompt enter "linux mem=exactmap mem=256M@0" 3. follow any prompts to install system. error will result when drive is formatted... other errors will result of the kernel parameter is not included. (insufficient memory or error mounting root fs) Actual Results: Crashed on Install. Expected Results: Expected 7.1 to install.
call trace: c01318cc c0116a79 c01301ce c0120ccf c0120d9d c0140870 c01330d1 c0133495 c013356f
call trace: c01318cc c0116a79 c01301ce c0120ccf c0120d9d c0140870 c01330d1 c0133495 c013356f c012de96 c01335a3 c0108eaf
Did you burn these cds your self?
Yes, I burned the CDROM's and created the floppies. Asside from the floppy disks booting much slower than the CDROM the results are exactly the same. boot with no kernel parameters: only 16 meg detected... 7.1 refuses to install claiming "not enough memory in this system" boot with "linux mem=256M" " loading initrd.img . . . initrd extends beyond end of memory (0x0ffffb27 > 0x01000000) . . . kernel panic: VFS unable to mound root fs on 08:41 I have been working with Matt Wilson who is of the opinion something is wrong with how the APIC table is reporting information. He sent me a custom boot disk that indicates that there is no information reported... I have verified that the system setup software from compaq is set for linux and I have tried all three options for the APIC setting (full table - mapped, full table, and disabled) with no change in functionality. If you believe the media is corrupt I will gladly download another copy of the boot image for the floppy and perform all of the tests I have done thus far again. If you honestly believe I will get different results with purchased commercial media I will start the government PO process to purchase the retail product... (usually a long drawn out purchase process due to redhat being a sole source for redhat linux...)
Actually, this sounds like more of a kernel problem than anaconda. Changing component.
Please try using 255 instead of 256. The bios tends to "steal" 1Mb of memory
no effect. nor does removing memory down to 64M and trying from there. problem is consistent with each of the 4 dimm (should not be physical memory problem) I just tried a 7.0 install followed by a 7.1 upgrade.... things went reasonably well until I got an pmd not installed Oops: 0002 . . . I'm figuring at this point that rh 7.1 on a proliant 2500 is not happening in the short term... I'll set asside one of these systems for further testing and build the others with 7.0 so I can work on documenting the configuration (trying to keep pace with the deployment schedule) still open to suggestions for makign 7.1 work...
I get the same errors with the memory. Compaq Prosignia 500 with 208M and only sees 16M. I installed 7.0 and appended RAM with no plms. 7.1 kernel panics with the "append=root". I have a guy in England with the same problem so its not your cd.
Bug 36940 has the commandline option for specifying how much memory you have: "mem=exactmap mem=0x4000000@0" would be 64Mb
Does this differ from "mem=exactmap mem=64M@0" ?
The lilo I tried it on didn't like the 64M, but if it works, it's the same.
Notice the proposed memory definition line in #36940 It seems you must point out that the memory from 640K to 1M is not regular memory.
*** This bug has been marked as a duplicate of 36940 ***
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.