El 21/11/2012, a les 11:06, Sampsa Laine <sampsa at mac.com> va escriure:
Really? How? I thought the level 3 works as a point to point link as I said, the level 2 stuff would let things like DECNET run.
Got a sample say Debian/390 setup you'd like to share?
Oh, I think so... but I'll have to dig for it. I know it works for z/OS (I can tn3270/ssh into the emulated machine, use IMS Connect, access DB2 remotely, ) so there is no reason for it to not work for Linux.
As for level 2, I think the Hercules implementation uses tunX devices, which work at level 3, so I guess it simply won't work for anything not IP.
I'd like it to support other protocols encapsulation... XCF for instance. Hercules-based Parallel SYSPLEXes would be doable then :)
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
On 21 Nov 2012, at 12:01, Jordi Guillaumes i Pons wrote:
El 21/11/2012, a les 10:31, Sampsa Laine <sampsa at mac.com> va escriure:
z/OS, AIX/ESA[370] and Debian/390x all under one emulated z/Arch box. That WOULD be cool.
And slow. VERY slow :)
I wasn't thinking of using this for production use. Compiling the kernel on Hercules was a giggle, took like 2 days.
If only they could get Ethernet support working (right now the tun/tap thing is a weird point-to-point link AFAIK).
It is. It works fine for TCP/IP (at level 3) but no level 2 so far.
Really? How? I thought the level 3 works as a point to point link as I said, the level 2 stuff would let things like DECNET run.
Got a sample say Debian/390 setup you'd like to share?
Sampsa
El 21/11/2012, a les 10:31, Sampsa Laine <sampsa at mac.com> va escriure:
z/OS, AIX/ESA[370] and Debian/390x all under one emulated z/Arch box. That WOULD be cool.
And slow. VERY slow :)
If only they could get Ethernet support working (right now the tun/tap thing is a weird point-to-point link AFAIK).
It is. It works fine for TCP/IP (at level 3) but no level 2 so far.
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
On 20 Nov 2012, at 18:27, Clem Cole wrote:
Be careful when you disparage it. AIX/370 could do some good stuff and I know of a number of firms swore by it for their engineering development teams. One large semi-conductor firm used it for all the simulation of an extremely large processor in the 1980s. At the time, they claimed no other system could handle it. In fact because of the support for process migration and rolling upgrades they keep the simulation running across some upgrades to the system.
Right I want a copy - I could run this under VM in Hercules, right?
z/OS, AIX/ESA[370] and Debian/390x all under one emulated z/Arch box. That WOULD be cool.
If only they could get Ethernet support working (right now the tun/tap thing is a weird point-to-point link AFAIK).
Sampsa
On 11/19/2012 09:18 PM, John Wilson wrote:
It's just such an amazingly bad match for that type of hardware
architecture. *shudder*
Yeah we had it at RPI and it was insanely slow. Sometimes it would drop
you during login because it couldn't process your username/password within
the 60-second timeout. And the software BS didn't help. The C compiler
was High-C and it had all kinds of bugs including choking on its own
#include files.
Fear.
Another side of the same mainframe ran MTS (from UMich) and a lot of people
never got over the culture shock, but it sure ran like a bat outta hell!
So it wasn't a slow computer, unless they horribly misconfigured the VM
AIX/370 was in, I guess.
VM tuning is a bit of a black art, even more involved than VMS tuning. (but VM doesn't have autogen ;)) Given the penchant for secrecy in the IBM world, at least compared to the DEC and UNIX worlds, I guess it's pretty conceivable that such things could happen.
P.S. I wrote the ROM half of PD.SYS for E11's new PDT-11/150 emulation and
now I'm working on the PDT-11/130 (TU58 with mutated DLV11) version for
good measure ... but slightly sidetracked attempting Terak stuff. I guess
this release is going to be all about the weird LSI-11-based micros!
That's gonna be just too much fun. :-)
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
On 11/20/2012 03:32 PM, Gregg Levine wrote:
And speaking of that, there are materials available on the TUHS ftp
site for it.......
For AIX?
Hello!
**Suddenly something resembling a crashing and demolishing sound is
heard outside.**
No for Ultrix. And for the VAX.
Oh. Yes, of course. :)
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
On Tue, Nov 20, 2012 at 12:22 PM, Dave McGuire <mcguire at neurotica.com> wrote:
On 11/20/2012 10:23 AM, Gregg Levine wrote:
And speaking of that, there are materials available on the TUHS ftp
site for it.......
For AIX?
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Hello!
**Suddenly something resembling a crashing and demolishing sound is
heard outside.**
No for Ultrix. And for the VAX.
-----
Gregg C Levine gregg.drwho8 at gmail.com
"This signature fought the Time Wars, time and again."
On 11/20/2012 10:23 AM, Gregg Levine wrote:
And speaking of that, there are materials available on the TUHS ftp
site for it.......
For AIX?
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
On 11/20/2012 09:28 AM, Sridhar Ayengar wrote:
Well I could use a copy of AIX370....
If you could, you'd be the first! What a waste of a perfectly good
mainframe.
I have to agree. I love UNIX, but running anything other than MVS on
something like that, or maybe VM, is just a sin.
Why not as a VM guest?
Well I thought that was sorta implied. :)
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Be careful when you disparage it. AIX/370 could do some good stuff and I know of a number of firms swore by it for their engineering development teams. One large semi-conductor firm used it for all the simulation of an extremely large processor in the 1980s. At the time, they claimed no other system could handle it. In fact because of the support for process migration and rolling upgrades they keep the simulation running across some upgrades to the system.
BTW; AIX/370 and AIX/PS2 are actually similar (much in common) code base. They were developed by Locus Computing Corp for IBM by the late Gerry Popek. Check out his and Bruce Walker's book: http://www.amazon.com/Distributed-System-Architecture-Computer-Systems/dp/0… for the details.
The really cool thing that these versions of AIX supplied was TCF - the Transparent Computing Facility. A total of 32 nodes of 370s and PS/2 could be "clustered" into a "single system image" and operated as a single computing environment (yes root on a PS/2 gave you root on the 370). I once saw some one unbox and plug a brand new PS2 into the network. He put a boot floppy in it, turn it on and with that have it "join" the cluster. In the background it populated its disk using replication.
Although neither VMS nor Tru64 would never completely match everything TCF could do, (they never had the full process migration stuff), Locus did sell some of the technology to DEC and it would land make it into TruClusters and DEC would develop similar things for VMS.
Check out: www.openSSI.org The technology still is around, although Beowulf style clusters (which are not full SSI) have usurped the market. Which is actually real shame. AIX/370, VMS, and TruClusters were really ahead of anything we have practically speaking today. I miss full SSI and a cluster where I can move process anywhere - I just saw one giant resource.
Clem
On Tue, Nov 20, 2012 at 10:23 AM, Gregg Levine <gregg.drwho8 at gmail.com> wrote:
On Tue, Nov 20, 2012 at 10:14 AM, Mark Benson <md.benson at gmail.com> wrote:
> According to Wikipedia, AIX/ESA will run as host or guest on System/370... if I read it right.
>
> --
> Mark Benson
>
> http://markbenson.org/blog
> http://twitter.com/MDBenson
>
> On 20 Nov 2012, at 14:28, Sridhar Ayengar <ploopster at gmail.com> wrote:
>
>> Dave McGuire wrote:
>>>>> Well I could use a copy of AIX370....
>>>>
>>>> If you could, you'd be the first! What a waste of a perfectly good mainframe.
>>>
>>> I have to agree. I love UNIX, but running anything other than MVS on
>>> something like that, or maybe VM, is just a sin.
>>
>> Why not as a VM guest?
>>
>> Peace... Sridhar
>>
Hello!
Close but no noxious smoking device Mark.
That was AIX/ESA. I'm talking about AIX/370. And according to one
associate on another two lists, (One I manage.) that one ran on one
specific customer's machines in the same area that one of our
associates is based in.
If John Wilson is correct and it also ran on the mainframe there,
sharing space with the MTS setup then it adds to the mysteries behind
the product.
And to add the the amusement I also met a family of machines running
ESA grade operating systems, a good long time ago, also a crowd of
RS/6000 machines and a very unhappy IBMer. The surprised look the
woman gave me was worth it.
This was the same time when Ultrix was announced for the VAX, and then
work was being started on the Alpha family, especially since I saw a
couple of MIPS based workstations then as well.
And speaking of that, there are materials available on the TUHS ftp
site for it.......
-----
Gregg C Levine gregg.drwho8 at gmail.com
"This signature fought the Time Wars, time and again."
