Dave Beck (dave@arginine.umdnj.edu)
Mon, 9 Nov 1998 06:32:02 -0500
The swap limitation is a (poor) design decision made by Linus in the
x86 VM code. There are limits such that it was vastly easier to make
swap limited to 128mb than it was to build things correctly to be
unlimited. In a desire to keep the code the same on varrying platforms,
PPC code has the same limitation. The obvious solution is to simply
make your swap partitions in increments of 128mb... I have 4 at the
moment on my dual PII system for a total of 512mb. It is very clumsy,
but it does force you to think about how you lay your swap filesystems
out across disks and how you assign priorities. I must say that when
I dropped each swap partition onto a different disk and various controllers
I saw a significant speed increase in swap usage by assigning priorities to
fask disks and fast controllers.
Quoting Joel M Snyder (Joel.Snyder@Opus1.COM):
> Hi there.
>
> I'm continuing my venture trying to make the ANS 500/700 work properly
> with LinuxPPC. (Thanks to Nikos for suggesting getting a new kernel;
> the .126 kernel blew chunks, but I found a .125 kernel at samba which
> built and worked very well)
>
> I'm up against two problems now:
>
> 1) When trying to move the system from an external SCSI adapter to one
> of the internal IBM drives, I cannot seem to get the boot interface
> correct. In particular, I have copied all the data over, done a quik
> and mkswap (what's the story with this maximum 130 Mb swap space???),
> but during boot, second.b comes up from the new disk but cannot deal
> with it:
>
> Fatal error: Cannot read driver descriptor
> Fatal error: Unable to open filesystem
>
> (this is when I've changed boot-device from scsi/sd:> (this is when I've changed boot-device from scsi/sd:1@0 to
> scsi-int/sd:> scsi-int/sd:2@0).
>
> >From the error messages, I am guessing that it is getting through first.b
> and second.b, but it cannot deal with everything else.
>
> If I leave the original build drive on the external adapter, it finds
> it, boots, and jumps to the OS loaded on the internal drive just fine.
>
> When I built the kernel, I did not do anything special with "initrd," which
> my Red Hat book seems to babble about in regards to SCSI (egads their
> documentation is vile in places). Is this relevant for LinuxPPC?
>
> 2) One of the disks is a Seagate 18 Gb drive, the ST118273LW. Linux
> senses the drive type very quickly, but it cannot do the READ CAPACITY
> on the drive. The error that comes back is something like this:
>
> sdc: Spinning up disk...<6>ncr53c825a-1-<6,*>: FAST-10 WIDE SCSI 20.0 MB/s (100
> ns, offset 8)
> ...........................................................
> ......................not responding...
> sdc : READ CAPACITY failed.
> sdc : status = 1, message = 00, host = 0, driver = 28
> sdc : extended sense code = 2
> sdc : block size assumed to be 512 bytes, disk size 1 GB.
>
> and then later
>
> sdc: SCSI disk drror : host 1 channel 0 id 6 lun 0 return code = 28000002
> extra data not valid Current error sd08:20: sense key Not Ready
> Additional sense indicates Logical unit not ready, cause not reportable
> scsidisk I/O error: dev 08:20, sector 0
> unable to read partition table
> (no surprise)
>
> I can probably deal with (2) by failing back to a smaller drive (although
> that's not why this machine exists in the first place), but if I can't
> get it to boot&run from the internal drive, I'm going to have to just give
> up the whole project.
>
> Any ideas/suggestions/etc will be much appreciated,
>
> jms
>
>
> Joel M Snyder, 1404 East Lind Road, Tucson, AZ, 85719
> Phone: +1 520 324 0494 (voice) +1 520 324 0495 (FAX)
> jms@Opus1.COM http://www.opus1.com/jms Opus One
-- Dave Beck dave@arginine.umdnj.edu "DNA is the ultimate assembly Computer Science and Biology language for the ultimate Drexel University, Philadelphia PA computer: life"
This archive was generated by hypermail 2.0b3 on Thu Dec 10 1998 - 16:55:54 PST