DFly fdisk issues continued... (fwd)

Chen Zhao dfly at nixwiz.com
Wed Jul 14 03:57:52 PDT 2004


Hi, I see I'm not the only one to get hit hard with this bug.
I mentioned the strange behavior of my installed OSes to
cpressey and ended up doing nothing but reinstalling XP, DFly, 
and FBSD-Current over and over again for the last two days.
Ended up thinking it was a quirk in XP until I decided to do
away with the XP bit.

I'm including the installer log I sent him so maybe others
can take a look at it too.  I guess this would be a disk
geometry issue in DFly's fdisk?

----- Forwarded message from Chen Zhao <dfly at xxxxxxxxxx> -----

Date: Wed, 14 Jul 2004 02:00:40 -0400
From: Chen Zhao <dfly at xxxxxxxxxx>
Subject: DFly fdisk issues continued...

Well, right after I gave up on tyring to put WinXP on the same disk,
I ran into something else rather odd, which probably contributed to
the problems I had before.  Now, WinXP is no longer involved and
so cannot be blamed...

I sliced the same 180gig drive into 4: 100gig, 40gig, 20gig,
and whatever is left over, around 15gig.

100gig slice partitioned as follows:
ad0s1a: 256M /
ad0s1b: 4096M
ad0s1e: 4096M /var
ad0s1f: rest /usr

I install DFly on the first slice, and everything went smoothly, 
until I tried to compile some ports and ran out of disk space.  
WTF?

Actually looking at my partition set up after the install for the 
first time, I see I am missing massive amounts of space.  Dfly fdisk 
shows that the first slice isn't starting near the beginning, but 
rather in the middle of where it was supposed to start.

I'm attaching the new install log.  It's not complete as I exited
after the fdisk section.  It shows the truncation of the slice.
Seems DFly and FreeBSD-Current fdisk don't calculate disk geometry
the same way or something...

The internal routines show the corret info, but fdisk isn't using
it.

Installer started
-----------------
+ Creating DFUI connection on ``9999''
+ Connecting on ``9999''
+ Starting installer state machine
,-<<< Executing `[internal] maybe_makedev()'
| Seeing if ad0 needs to be created...
| No, ad0 already exists.
`->>> Exit status: 0

,-<<< Executing `fdisk -s ad0 >/tmp/temp.fdisk || echo '' >/tmp/temp.fdisk'
`->>> Exit status: 0

,-<<< Executing `[internal] discover_slices()'
| Found slice #1, sysid 165, start 63, size 209712447
| Found slice #2, sysid 165, start 209712510, size 83875365
| Found slice #3, sysid 165, start 293587875, size 41929650
| Found slice #4, sysid 165, start 335517525, size 26362665
`->>> Exit status: 0

,-<<< Executing `rm -f /tmp/temp.fdisk'
`->>> Exit status: 0

,-<<< Executing `[internal] maybe_makedev()'
| Seeing if ad1 needs to be created...
| No, ad1 already exists.
`->>> Exit status: 0

,-<<< Executing `fdisk -s ad1 >/tmp/temp.fdisk || echo '' >/tmp/temp.fdisk'
`->>> Exit status: 0

,-<<< Executing `[internal] discover_slices()'
| Found slice #1, sysid 131, start 63, size 144522
| Found slice #2, sysid 130, start 144585, size 1012095
| Found slice #3, sysid 131, start 1156680, size 37945530
`->>> Exit status: 0

,-<<< Executing `rm -f /tmp/temp.fdisk'
`->>> Exit status: 0

,-<<< Executing `[internal] maybe_makedev()'
| Seeing if ad0s1 needs to be created...
| No, ad0s1 already exists.
`->>> Exit status: 0

,-<<< Executing `fdisk -s -1 ad0 >/tmp/slice.fdisk'
`->>> Exit status: 0

,-<<< Executing `[internal] set_slice_sysid_dfly()'
| Creating fdisk script to set sysid of partition 1 to 165...
|--/dev/ad0: 22526 cyl 255 hd 63 sec
|--Part        Start        Size Type Flags
|--   1:          63   209712447 0xa5 0x00
|--   2:   209712510    83875365 0xa5 0x80
|--   3:   293587875    41929650 0xa5 0x00
|--   4:   335517525    26362665 0xa5 0x00
| Cylinders: 22526, Heads: 255, Sectors/Track: 63
| Start: 335517525, Size: 26362665, Type: 165, Flags: 0
`->>> Exit status: 0

,-<<< Executing `fdisk -v -f /tmp/new.fdisk ad0'
| fdisk: WARNING line 1: number of cylinders (22526) may be out-of-range
|     (must be within 1-1024 for normal BIOS operation, unless the entire disk
|     is dedicated to DragonFly)
| ******* Working on device /dev/ad0 *******
| parameters extracted from in-core disklabel are:
| cylinders=22526 heads=255 sectors/track=63 (16065 blks/cyl)
| 
| Figures below won't work with BIOS for partitions not in cyl 1
| parameters to be used for BIOS calculations are:
| cylinders=22526 heads=255 sectors/track=63 (16065 blks/cyl)
| 
| Information from DOS bootblock is:
| 1: sysid 165,(DragonFly/FreeBSD/NetBSD/386BSD)
|     start 335517525, size 26362665 (12872 Meg), flag 0
| 	beg: cyl 405/ head 0/ sector 1;
| 	end: cyl 1021/ head 254/ sector 63
| 2: sysid 165,(DragonFly/FreeBSD/NetBSD/386BSD)
|     start 209712510, size 83875365 (40954 Meg), flag 80 (active)
| 	beg: cyl 1023/ head 255/ sector 63;
| 	end: cyl 1023/ head 254/ sector 63
| 3: sysid 165,(DragonFly/FreeBSD/NetBSD/386BSD)
|     start 293587875, size 41929650 (20473 Meg), flag 0
| 	beg: cyl 1023/ head 255/ sector 63;
| 	end: cyl 1023/ head 254/ sector 63
| 4: sysid 165,(DragonFly/FreeBSD/NetBSD/386BSD)
|     start 335517525, size 26362665 (12872 Meg), flag 0
| 	beg: cyl 1023/ head 255/ sector 63;
| 	end: cyl 1023/ head 254/ sector 63
`->>> Exit status: 0

,-<<< Executing `rm -f /tmp/slice.fdisk'
`->>> Exit status: 0

,-<<< Executing `rm -f /tmp/new.fdisk'
`->>> Exit status: 0

,-<<< Executing `disklabel -B -r -w ad0s1 auto'
`->>> Exit status: 0

,-<<< Executing `rm -f /tmp/install.disklabel.ad0s1'
`->>> Exit status: 0


----- End forwarded message -----





More information about the Bugs mailing list