The collected works of jpstewart - Page 3

Wow, that's neat to hear! I've got an A2500/UX in pieces* waiting to be assembled just so that I can play around with Amix. Now it seems I may not need the actual hardware.

(* actually it's a standard A2000 that I bought new in 1989 and used for more than 5 years. I recently pulled it out of storage and bought the add-ons to upgrade it to the equivalent of an A2500/UX. I haven't had a chance to put everything together yet.)

mia wrote: This makes me think, does this also mean that netbsd/openbsd and perhaps m68k linux can now work on winuae?

According to winuae.net , yes! Under "New Features" for WinUAE 2.6.0 it says: "Full 68030, 68040 and 68060 MMU emulation. Amix, Linux, NetBSD, Enforcer, WHDLoad MMU option and more fully working."

Nice!

Call me crazy, but I like the older/slower systems better for some reason. Maybe it's just because you could do so much with so little hardware back then, before software bloat got us to where we are today. And the programmers who got the software to do so much in so little space get my respect, too.

I have an R4K Indigo that I was going to install IRIX 4.0.5 on until I discovered my CD is completely corrupt. I'd actually prefer to have gotten a R3K instead (in part because it's even older; in part becuase I already have an R4K system, my Indigo2). And you say you have the beige keyboard and old monitor. That makes for an impressively complete old system.

So from my crazy, backwards, retro-focused viewpoint that is one very nice machine you have there!

That's a nice photo. I love the monitor. I don't think I've ever seen one of the old beige monitors before, only the granite models.

Enjoy that setup!

recondas wrote: Back from a road trip along the east coast into the Canadian Maritimes. In chronological order the route went through NC, VA, WVA, MD, PA, NJ, NY, CT, MA, NH, ME, NB, PEI, NS, NB, ME, NH, VT, NY, PA, MD, WVA, VA, and back to NC.

Wow! That sounds like a nice, but very long, trip. Mind if I ask how many days you were on the road and/or how many miles you covered?

It sure looks like you got some spectacular photos out of it, too!

recondas wrote: Twelve days, four thousand miles.

That's a lot of mileage for only twelve days. I'd be exhausted trying to cover half that distance in that amount of time (by car; motorcycles aren't my thing). Maybe the added dose of fresh air on the motorcycle helps you stay awake!

BetXen wrote: On the OS side, it is for the moment running 6.5.22, but is very slow.

If you want to stick with 6.5.22, upgrade your RAM! The R4K Indigo can handle up to 384MB, and (in my experience) anything under 128MB is just too little for 6.5.x. 192MB and up is more realistic for actual use. I was quite pleased with the performance boost I got on my Indigo2 when going from 64MB to 320MB. (I realize it's not the same machine, but still....)

Of course, if you prefer the older IRIX versions anyway, then there's probably no need to upgrade RAM.

blowout6 wrote: Also, I'd imagine there should be driving / racing simulation software.

I wonder how feasible it would be to get TORCS running on IRIX. I suspect it would be better on an Onyx than anything in my humble collection, if it could run on IRIX at all.

Probably any simulator (even driving ones) will be like the high-end flight simulators the other posters have been discussing: they'll be proprietary software and may require specialized hardware (in addition to the IRIX machine).

jan-jaap wrote: Strangely, the airflow is back-to-front. Ideally I should mount it in the back of the rack, but that's not very convenient.

That seems to be somewhat common on switches. I guess in a data centre environment it makes sense to mount the switch at the back of the rack. Putting the switch ports at the back where the servers' ports are can simplify cabling --- no need to pull cables through to the front of the rack. I'll have to check the airflow on my 4100s now. I never thought about that until you mentioned it.

jan-japp wrote: As a bonus, it also come with Extended Fabric, Fabric Watch, Performance Monitor and Trunking licenses. That should keep me busy for a while

Congratulations on a nice score! Those licenses cost a fortune on their own. Have fun with it.

I've had an Indigo for about a year and half now. Apparently I never properly introduced him to the forum.

At the moment, all I can do is get the PROM-level hinv info. This machine was going to be my IRIX 4.0.5 playground, but I discovered that my 4.0.5 CD is damaged. So the machine is still waiting for an operating system. I'm thinking about NetBSD instead. Certainly not IRIX 6.5 since my other SGIs are already running that much faster than this little guy could!


The tape is an original SGI DDS/DAT drive, but other than that there's nothing special about this Indigo. I replaced the TOD clock battery a little over a year ago, so no troubles on that front.

I might consider selling/trading this system for an Indigo R3000 Elan instead. I don't have an R3K system of any sort and would like one so that I can compare it to the faster machines. Especially since my Indigo2 has a 250MHz R4400 and that's a much faster specimen from the R4K family. And I just have this weird fascination with older/slower hardware.

vishnu wrote: I've got Sun Studio 11 and Sparkworks 1.129 on my Blade 2500 but since it's 64 bit there's no point trying to compile it on that?

Just pass the "-m32" option to the compiler to compile in 32-bit mode. So either set CC="cc -m32", or set CC="cc" and CFLAGS="-m32", or similar. There are plenty of ways to compile 32-bit code on a 64-bit machine.

vishnu wrote: I'm getting system header errors, could somebody kindly do me the favor of checking what version of /usr/include/inttypes.h you have? Mine says:

Module: inttypes.h
$Revision: 1.4 $
$Date: 1998/06/23 00:38:11 $

It's failing with:

Code: Select all

cc-1020 CC: ERROR File = /usr/include/inttypes.h, Line = 188
The identifier "va_list" is undefined.

va_list is defined in varargs.h but that won't compile either:

va_list is also defined in stdarg.h which is included by inttypes.h (at least on my system). So I'd suggest checking:

1) Does your version of inttypes.h actually include stdarg.h before any mention of va_list?
2) Does your version of stdarg.h have typedef char *va_list somewhere?

I'll also say thanks for the warning.

I'll have to put "inspect SS20 PSU" near the top of my "to do" list. Now that I know about the potential for trouble lurking in there, I'd certainly prefer to fix the PSU before any serious problems crop up or any damage is done.

That was a great tip to share.

duck wrote:

Code: Select all

knightrider% date && uptime
Sat Jan 25 17:19:51 EET 2014
5:19PM  up 507 days, 21:10, 5 users, load averages: 0.15, 0.35, 0.38


No UPS .

I want your power company! That's really impressive.

I'm hard pressed to get uptimes like that with a UPS. I've got 20 minutes of battery power but around here the power tends to go out for less than a minute or for several hours. There's no middle ground. The brief power flickers happen too often for my tastes. The major outages happen every year or two, it seems. The last one (over 4 hours) was just 34 days ago. Although at that time, people in nearby cities were in the dark for the better part of a week thanks to an ice storm. So I got off lucky.

I did once have a trio of Dell PowerEdge 1750 servers running Linux all make it (just) past 800 days of uptime before the power went out for several hours. That was my best uptime but can't compare to anything I've seen a long while. And almost always due to power outages.

Well, I inspected my SS20's PSU last night. Fortunately there are no signs of leaking electrolytes...yet. Which is a good thing, because looking at it suggests that re-capping it is going to be beyond my (current) soldering skill level. Those daughter boards are going to be tricky to work with for sure. Here's hoping mine keeps working until I get more comfortable with soldering circuit boards! (I'm definitely a software guy.)

Again, thanks for the warning. I wouldn't have thought to inspect my PSU without it. And even bigger thanks for the very detailed replacement parts list. It looks like a lot of work must have gone into finding suitable replacements for all those capacitors. It will make it a lot easier for the rest of us to fix ours when (not if!) the time comes.

Nice! I've long wanted an ANS just because of how rare and obscure they are. (And they run Unix!) Since that's unlikely to ever actually happen, I enjoyed reading about yours.

SAQ wrote: If you want an Apple product that runs UNIX get A/UX

Oh, that's on the wish list, too! (And more likely to actually happen.) It's the rarity of the ANS line that I find appealing. (The "what the heck is that?" factor, I guess.) That same rarity is what makes ClassicHasClass' write-up and pictures so interesting.

I recently won an ebay auction for a beige Indigo keyboard and mouse set which prompted me to try to play with this machine again. I tried to install NetBSD to no avail. It wouldn't read the CD I'd burned NetBSD onto. That led me to suspect that maybe it was the drive (an HP DVD-ROM in an external enclosure that I also use with my Octane) causing my earlier problems. So I tried a known-good 6.5.18 installation tools CD. Again, it wouldn't read the CD.

So I pulled the HP DVD-ROM from the external case and dropped in an old Plextor CD-RW that had recently been "retired" from another machine. Lo and behold, that drive read any CD I threw at it and the Indigo.

Since there was nothing wrong with my 4.0.5F CD, I went ahead and installed that:

uname -a:


hinv -v:


/usr/gfx/gfxinfo -v


The good old DAT drive in the Indigo happily ran through a full backup using IRIX's backup manager. I didn't expect that to work so well on the first try, considering the age of the drive. I'll have to get networking set up soon, but I'm having plenty of fun with the system as it is!

Until today I'd only ever used IRIX 6.5.x, nothing older. I'm actually quite surprised at how nice IRIX 4.0.5 is from an end-user standpoint compared to SunOS 4.1.4 considering they're of similar vintage (AFAIK). I haven't tried any programming on it yet, but I suspect that will be an "adventure" if/when I do.

Here of some photos of what it looked like when I got it. All I've had to do cosmetically was clean the dust off of it!

foetz wrote: nice machine and a good call running irix there. doesn't make much sense running anything else.

I agree. I was only considering *BSD because I thought my 4.0.5 CD was bad and didn't want to go with 6.5.x with so little RAM. I was very happy to find that with a different CD drive I was able to install 4.0.5.

foetz wrote: 4 is also interesting from a dev perspective because that was the coff era and the compiler doesn't know a lot yet

I had forgotten that the IDO was a seperately licensed product (unlike the Developer Foundation CDs that come with the base 6.5 install set) so even getting GCC (up to 3.x is supported AFAIK) runing will be a challenge. But that's another project for another day. I'm still very excited to have the little guy up and running.

ClassicHasClass wrote: That /is/ a very nice classic Indigo.

Thanks! That means a lot coming from someone with as nice a collection as yours is! Although when you call it a " classic Indigo" it makes me affraid you're going to try to claim it as your own!

Thanks for the comments guys!

Wow, that's fascinating info. Thanks for sharing!

I'm shocked at how little disk space 4.0.5F takes up and yet is still so full-featured. After a full install (and I do mean full --- I simply did 'install *' --- I'll figure out what to remove later), I get:


Install ing the full IDO (including C, C++, Fortran77, and Pascal compilers), it gets up to:


I'm sure there's plenty of stuff that I'll never use that could be trimmed out of there, too! Considering that I'm typing this on a (four year old) Linux box with 12GB of RAM , seeing a full install with developer option using less than 0.4GB of disk is rather impressive. (People wonder why I like playing with old hardware and software. It's impressively effecient stuff!)

IRIX 4.0.5 looks mighty similar to 6.5.x, too. The toolchest is rather familiar, but not completely identical. There's no Icon Catalog and no icons on the desktop itself, but there is the WorkSpace window which combines several features of both with a bit of file management on top.


Scrolling down to the bottom of the WorkSpace window shows icons for some image tools. The imgsnap icon is highlighted in pale yellow as a result of me using it to take the screenshot.


And last but not least, winterm shell windows haven't changed much over the years.


The filesystem layout sure is different. /var is missing, for example! Most of what I'd expected to see there is under /usr so that's no big deal now that I know where to look. The installer defaulted to making / only 8MB but that was too small, so I doubled it. Other than that, it's been pretty smooth sailing with 4.0.5.

This old boy has quickly become my favourite system in the collection! At least until I try compiling stuff.... The warnings in this thread from foetz are rather foreboding.

jan-jaap wrote: How something can be airmail and still take 3 weeks to arrive is a little beyond me

I hope you've got a tracking number for that package. Watching a package progress through the postal/courier system can be downright entertaining. I just received a package that spent 8 days making a journey that Google Maps says you can drive in 7 hours. I know postal trucks aren't the fastest things on the road, but still! I've also received courier packages from the United States to me in Canada that travelled US->Canada->US->Canada. And its first trip to Canada landed 1500km from here for no apparent reason. It made me think they just randomly flung the package at Canada, missed the mark, and tried again.

Tracking international packages can be fun, as long as you don't expect a logical or effecient path.

hamei wrote: The post offices have international agreements where the price is set by the sending country, but do they keep track ? Do they balance how much each country sends versus how much it receives ? Or do they just say "To heck with it" and let it all balance out in the wash ?

The last couple of paragraphs in the "Overview" section of the Wikipedia article about the Universal Postal Union addresses that. In short, there are fees these days.

Wow!

That's a very detailed write-up, foetz. Well done. I particularly liked the fact that you didn't just review the software but also provided a fair bit of history in the intro. It's clear you've put a lot of work into this. Thanks for writing it!

Wow! Congrats on finding such an awesome Amiga. I know how rare the A4000Ts are, and then 060 and upgraded graphics, too! You must be super excited. I'm sure you'll have tons of fun with it...more than enough to justify its price tag.

What's the status of the RAID component devices look like? Is the whole array showing as failed because too many component devices are thought to be failed? If so, does the management interface let you manually/forcibly re-enable the component devices?

I have zero experience on POWER systems or their RAID controllers. But I used to have an IBM ServeRAID 6M in a PC server. Just about any unclean shutdown of the system (but particularly those that happened as a result of bad RAM) would cause the 6M to think that two devices in its RAID-5 array had died. Booting from the ServeRAID CD and using its interface to change the drives' status from "Faulty" back to "On-line" was all that was needed. The data on the disk(s) was all there so the controller didn't re-sync/re-build the array. So I had no data loss at all. In my case, it was just a confused controller but a healthy array.

Again, I have no idea how your POWER system compares to my ServeRAID 6M, but your situation (hardware failure leading to a crash leading to a hosed array) sounds just like what I experienced. It might be worth a try if the tools allow it, and if you can accept the fact that it might not work at all.

Glad to hear it all worked out for you.

And sorry the info I posted wasn't useful. I had hoped that there might be some similarity between IBM's PC ServeRAID cards and the ones for their POWER systems. What I hadn't realized (and really should have ) was that your system was a generation or two newer (using SAS) than the old ServeRAID 6M (U320 SCSI) that I'd used. I hope I didn't waste too much of your time with my irrelevant ramblings.

vishnu wrote: If you do `make dist` from the top of the source tree it should rebuild the whole package into a cross platform tar.gz and excise all the GNU-isms from the makefiles. I say "should" because not everyone uses the auto tools the way they're intended

It would appear that Netsurf doesn't use autotools at all. They've got their own build system, AFAICT.

I had a quick look at it and its instructions. It doesn't look like Netsurf's QUICK-START instructions will work. That method relies heavily on their custom env.sh script which doesn't look good for IRIX. The Netsurf BUILDING-GTK instructions, however, do look viable. I'll keep working on it over the weekend and see how far I get. (Unless somebody else beats me to the finish line!)

So far I've seen some minor annoyances, but nothing terrible. It seems to assume GNU make and GNU install, for example. Their build system doesn't know about MIPS Pro (nor would I expect it to, really). Those shouldn't be show stoppers, though. (But who knows what lurks in the depths!)

With only a minor tweak to the Makefiles, I've built most of the included libraries with GCC on IRIX. I'm not 100% sure yet, but I think the ones that didn't build aren't strictly necessary anyway. I'm planning to work on the NetSurf core later today.

Building with MIPS Pro will be a different story. Their custom build system looks (at first glance) to be fairly portable. But then the Makefiles for the individual libraries go and add GCC-specific options to the compiler command line. Depending on how it goes with GCC, I may or may not attempt to build with MIPS Pro.

Details (and patches) to follow.