Been there too :( Really, boot a memtest86 kernel and test your hw. #emerge memtest86 Don't remember if there is any memtest-option on the LiveCD...?
Cheers! /C -----Original Message----- From: Kevin [mailto:[EMAIL PROTECTED] Sent: 9 januari 2005 17:10 To: [EMAIL PROTECTED] Subject: [gentoo-user] Problem with 2004.3-r1 universal live cd image Hi All- I'm having some installation problems that are associated with the subject CD image: install-x86-universal-2004.3-r1.iso I recently downloaded and burned this iso image to CD. I checked the gpg signature and md5sum before and after burning and they were both ok. On two very different hardware platforms, I've used this image to install Gentoo systems and either during or after installation (or both), I've seen these problems: Various "emerge -v X" commands have failed with segfaults and messages along the lines of, "...not reproducible so probably hardware or OS related...". In one case, even the tar -xvjpf stage3-... command segfaulted. I managed to get it to complete after several tries and by removing the "-v" flag (for some odd reason, the "-v" definitely has an impact on whether or not segfaults occur on both boxes with either tar or emerge). In this case (with the tar files), I md5sum'ed the stage3 bzipped tarfiles and they were ok, but bzip2 -vvt reported integrity problems (with the one I wanted and all the others on the CD also, except for one of them). Later, running bzip2 -vvt reported no integrity problems, so I went ahead with the tar -xjpf command without problem. Clearly, something is inherently unstable here because running the same command (bzip2 -vvt) on the same file stage3-whatever fails at one point in time and succeeds at another (no reboot in between). I've also seen kernel Oops and kernel panics. The two hardware platforms were: 1) An eMachines T2824 desktop computer. Specifications: -Intel Celeron D 325 Processor (2.53 GHz, 256k L2 Cache, 533 MHz FSB); -Intel 845GV Chipset -8-in-1 Digital Media Manager (USB 2.0, Secure Digital (SD), Smart Media, Compact Flash, Memory Stick, Memory Stick PRO, Micro Drive, Multimedia Card) -standard 256 MB DDR (PC 2100) RAM replaced with 1024 MB of same -http://www.emachines.com/support/support_info.html?prodName=T2824 for others if you're interested. I used the stage3-pentium4 tarfile for this. I'm really not sure if that's right or if I should be using one of the others (i686? pentium3? x86? With Celeron processors, how does one decide which CFLAGS and binaries to use?). I chose pentium4 because memtest86+ v1.11 reported the CPU as a pentium4 and I've seen some hints that some Celeron processors are "Pentium-4 flavors." 2) A very old Cyrix 6x86 with 132 MB RAM, using the stage3-x86 tarfile. I have run memtest86 and memtest86+ on both boxes (at least 10 passes each) and discovered no errors with the memory in either case. Interestingly (at least to me), memtest86+ v1.3 reports FSB at 133 MHz rather than 533 MHz. Could this be a clue as to the cause of the problems? Does anyone have any thoughts on what's causing this strange behavior? I installed first on the eMachines box and suspected hardware problems, but the segfaults were not so frequent that I could not proceed, so whenever I encountered one, I just ran the command again until it completed successfully and figured I could troubleshoot the hardware more later, after the base install. But now that I've seen the same sorts of problems on this Cyrix 6x86 CPU, I'm less inclined to think hardware problems and more inclined to think that there may be some problem with the software on the subject iso image. Has anyone else encountered this? Any ideas as to what else could be the problem if not the memory? As I said, the md5sum is fine on the cd image, the cd itself, and various tar files on the cd. Is there another hardware component that could be causing this (aside from the memory). TIA for any thoughts. -Kevin -- [email protected] mailing list -- [email protected] mailing list
