From: "Brian K. White" <br...@aljex.com> Subject: Re: Alternative to SCO Date: Fri, 5 Oct 2007 23:21:14 -0400 Message-ID: <00cf01c807c7$f39bf5f0$6800000a@venti> References: <firstname.lastname@example.org>
<email@example.com> > Actually, it's easier to deal with the old systems. The main > advantage is the very small sizes of the OS and applications. For > example, the Xenix system fits in about 500MBytes, most of which is > data. Just to point out for those unfamiliar with Xenix, the OS itself is actually under 50, heck, under 10!. An entire Xenix filesystem can not even exceed 512M, though you can have multiple 512M fs's mounted for a total system that exceeds 512M. However, a 320M Xenix system IS huge when you remember all the options you don't have for getting data off of a Xenix box. * Networking - nope. (it existed but I never saw a single box with it, just like the C compiler) * Serial (kermit/rz/sz not PPP, no networking!) - yes, but only at 9600 unless you have a 3rd party serial card driver and card and the right kind of serial cable. Hopefully the box has a 3.5" floppy drive at least to copy gzip, rz, sz, and kermit onto the box in the first place, or else you are in for some real nostalgia using cu and %put or %take and learning how great we all hove it now that most communications protocols have built-in error detection, correction and/or retransmit, which cu put/take does not! Oh, and beware even if it is a 3.5" drive, it might well actually be a 720K drive not a 1.44M drive! * hd transplant to new hardware to continue to run xenix - if you can find a 486-100 or slower box that actully runs, reliably. (think 20 year old dried out capacitors all out of spec making the circuit unstable at the very least, if not a fire hazard! Not kidding! Seen it! Caps out of spec, so voltage regulator circuit out of spec, cooked several parts of the board, dried up stickers on components and on circuit traces caught fire!). * hd transplant for linux to read - linux can mount xenix fs's but it can't parse xenix divvy tables to know where the fs's start & stop. Theoretically you can read the divvy table yourself while still in xenix and write down the numbers and calculate the offsets yourself manually and use dd in linux to extract the fs to a seperate file. If you can manage that then linux can loopback mount the extracted image files effortlessly. I never managed it but I think Bela once described the basic steps here. * hd to hd copy to create partitions or cpio files that linux CAN read - if you can find a hard drive small enough that an old 486-66 motherboard bios can even see it. * tar/cpio to floppies - yep, that works. You need to know how to reassemble sco's multi-volume tar files unless you happen to be restoring onto a newer sco os instead of linux or freebsd. It's not conceptually hard, but is incredibly tedious and thus prone to human error, but a simple script can and has been written to automate that. So the only problem is that it's almost impossible to write to 200 floppies and read from them and have not one of them turn out corrupt. An you must be ok with spending a few solid hours feeding/swapping floppies into machines every couple minutes. ...hm... I could modify the sco-tar script so that it will retry any given volume and not proceed on to subsequent volumes until you say this one was OK, giving you a chance to throw away the bad floppy and regenerate the volume on the source machine instead of having to start the entire job over from the beginning. But is there a way to do the same thing on the source machine that is generating the volumes? Have it regenerate a given volume as many times as you ask, before proceeding on the the next, giving you a chance to read a volume and discover it's bad instead of having to start the entire job from the beginning. hm... * tape drive - if you can find working media for the ancient tape drive in the xenix box, and if that drive and it's controller card can then be moved and be installed on linux, or if you can install a newer drive on the xenix box. * cd burner - put down the Lagavulin.
So, when faced with an old xenix box, you are actually faced with a pretty
big problem getting that system off onto other hardware or OS in less time
than a full weekend unless you are well prepred ahead of time with things
you would have a hard time finding on short notice.
Someone recently posted on this list that they worked out a recipe for
taking a full xenix raw hd image and running it in a bochs or qemu emulator.
I'm dying to try that. If that's possible, then that opens up a fast
migration option by just taking the hard drive out of the xenix box and
plugging into any linux box, dd an image of the whole disk in linux in no
time. You could do it with a laptop and a usb drive enclosure and the drive
could be cloned and right back into production in the old box in a few
minutes if necessary. Then you get that working in linux at your leisure,
then some other time update the clone and switch over to it for live
production in as little as a few minutes.
Brian K. White br...@aljex.com https://www.myspace.com/KEYofR
filePro BBx Linux SCO FreeBSD #callahans Satriani Filk!
Got something to add? Send me email.
One day my daughter came in, looked over my shoulder at some Perl 4 code, and said, "What is that, swearing?" (Larry Wall)