Le mer, 02 jun 1999, Konstantin Boldyshev a post� :
>
> I've experinced some weird problem with install program in 6.0 release. I do
> not
> know whether this is 2.2.9 kernel bug or install bug or both of them.. Here
> it is:
> I have two IDE hard drives - hda(WD 13GB) and hdb(WD 2GB), and also a CD as
> hdd.
> After kernel is loaded and install is started following kernel messages
> appear on 4 tty:
> hdb: lost interrupt
> hdb: drive_cmd: status=0x58 { DriveReady SeekComplete DataRequest }
> hdb: status_error
> hdb: drive not ready for command
> hdb: status_error
> hdb: drive not ready for command
> However if I interrupt install at this stage nothing bad happens, partiom
> tabel and MBR are OK.
> But if I proceed, and get to dialog with choice of Disk Druid and fdisk and
> choose Disk Druid - install tells me that partition table on hdb is broken
> and asks me whether to initialize it or not. And this is true - partition
> table on hdb is really broken at this time (first 4 sectors are zeroes!)
> However i can reinitialize it and proceed with installation and even install
> mandrake on hdb..
> This is true as for hard drive installation, I do not have burned cd yet..
> I've tried this several times - bug is reproducing.
> There was no such weirdness with RedHat 6.0 install. I run kernel 2.2.9
> (with uniform ide patch) several weeks on RedHat 6.0 and never expereinced
> anything like this. Fortunately I use hdb only for backup purpose and
> haven't lost any data.
> Anyone wants to try to reproduce this on his machine? :)
Yes, the problem is coming from few lines of code added in etc/rc.d/rc.sysinit
and called "Optimization of Hard drive" Cancelling these lines will cancel the
problem...
In fact datas are not lost but the hard drive does not run normally
MP
