Quote:
Originally Posted by daemonfowl
( but I still receive those 'wd0: soft error (corrected)
|
As mentioned in the original thread, that is an I/O error that is (eventually) recoverable. Its cause is unrelated to the segfaults described here. Possible causes were described to you in the original thread.
Quote:
|
..I chose bsd.rd and got this error..
|
The bootloader was unable to read the file from the drive. The bootloader is likely using BIOS I/O services to read the kernel from the drive; those routines may not provide recovery from soft (recoverable) I/O errors. This is conjecture, as I know nothing about PPC BIOS services on Macs or any other PPC based computer.
Quote:
|
does it mean iMac's Christmas festivities are over :-) ?
|
I don't consider an 'iMac' a single component. Surely there are parts that, if failing, could be replaced? It is possible the I/O problem is related to the drive or its on board electronics, or to the cable connecting the drive to the controller, or to the drive controller on the motherboard. As SMART did not log any cable communication problems, the I/O errors are more likely to be due to another component.
If this were my system, I would install 5.1-release to see if the segfault issue goes away. I would also (re)install sysutils/smartmontools and
run both a short test of the electronics and a long test of the media in the hopes that the I/O errors can be identified and, if media, perhaps even eliminated through sector reassignment. For forcing sector reassignment (after proof of a media issue) I use the "badblocks" program included in sysutils/e2fsprogs.
Quote:
|
(Note: those errors don't occur on ibook g3 running the latest snaphots.)
|
I cannot tell you (yet) why you are seeing the segfaults -- these are usually software and I have not (yet) seen any information to indicate otherwise. Your I/O errors are due to a hardware problem in electronics or cable or media, and short and long tests can prove or rule out two of them.
Quote:
A question off-topic :
When upgrading , do I have to use a respective new bsd.rd to boot from ? or does an old bsd.rd existing in / do the job ?
|
Take this question to another thread, please. But I would ask you to boot a ramdisk kernel for yourself and review the upgrade script and try to determine what it tests for and what it does. See if you can figure out how it knows it is looking for *52.tgz filesets and what will have to happen when -current shifts to *53.tgz filesets.