Port-alpha archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: writing cdhdtape to CD



On Fri, 5 Jun 2009, Ray Phillips wrote:

A quick test would be to pad the cdhdtape image to a multiple of 18K, burn that adjusted image, and see if it works:

dd if=cdhdtape of=/tmp/cdhdtape bs=18k conv=sync

That was a promising idea Michael. Executing that command produced a file 3778560 bytes in size. I burnt it to CD and tried to boot the PWS but the "block size is not 512 bytes" error occurred and the boot failed the same as usual.

That sounds like the problem of your CD drive using 2048 byte blocks, and the SRM on the PWS probably has to have 512 byte blocks.

I think the SRM on my CS20 can deal with the CD drive having 2048 byte blocks, and will do the buffering and addressing of 512 bytes when using the SRM read routine. I just successfully booted my CD created as above on my CS20.

% dd if=/dev/zero of=2048b bs=512 count=4
4+0 records in
4+0 records out
2048 bytes transferred in 0.001 secs (2048000 bytes/sec)
% ls -l 2048b
-rw-------  1 ray  wheel  2048 Jun  5 10:18 2048b
% cat 2048b >> cdhdtape
% ls -l cdhdtape
-rw-------  1 ray  wheel  3778560 Jun  5 10:19 cdhdtape
%

but apparently that idea was way off track. When I tried to boot the PWS using it the screen went blank and the machine rebooted as though it had been power-cycled.

Hmm, that should have created the same image as my dd command - you could compare the two files to see if there was any difference.

Fortunately I've been able to use CD-R/W media to minimize wastage. I've just used a couple of CD-Rs to ensure that wasn't a factor.

I'm not sure the CD drive in my CS20 would work with a CD-R/W. I know the old DEC drives that came with the vax, decstatsion, and early alphas didn't seem capable of reading CD-R/W media.

That's scary. Do you think booting with cdhdtape has affected your alpha somehow?

No, the machine has been rather tempermental before. After I upgraded to a 140GB drive on both my CS20 machines, it seems like they will boot up in some strange condition where the disk read speed is only 3-4 Mbytes/sec, and seems to corrupt the data while reading it off the disk. I first noticed it when fsck found lots of problems with the file system. After several different attempts at booting using different kernels, I can finally get it in a state where the disk speed is normal and reading doesn't corrupt the data and it will run fine. Then I tried running a 5.0 kernel on both CS20 systems, and they seemed to run fine until one of them crashed. After rebooting it, I had severe problems getting ntpd to sync and finally fingured out that the PCC source for the timecounter was getting configured with 1249660000 Hz instead of 833091856 Hz. If I boot with an old 4.0 kernel, or the 5.0 GENERIC.MP [which does not work with > 1 cpu currently], then my 5.0 GENERIC kernel gets the right frequency. Very weird. And yet another problem is that the CD drive seems to get in a state where it won't read anything until I eject and reload it [I ran into that last night when trying to test my lastes cdhdtape image, and had to wait until I got into my office to eject the CD].

--
Michael L. Hitch                        mhitch%montana.edu@localhost
Computer Consultant
Information Technology Center
Montana State University        Bozeman, MT     USA


Home | Main Index | Thread Index | Old Index