Off-Topic / Dumping Thread

[+mizapf]

13 hours ago, SteveB said:

The article was hilarious ... here is the Google Tranlation to English.

I, as a Java fan, found this article a bit too polemic, in particular the first part, which had a lot of factual errors. But the line

 

Quote

But what is JNDI? Jindi al Dap is the name of an ancient Arabic philosopher and mathematics pioneer who worked for Sun / Oracle to develop a system of directory lookups in Java [1]

 

made me smile. Let me try: جندي الدب

 

(Google Translate offers the translation "bear soldier" )

[SteveB]

"hilarious" meant I was very amused ... it does not mean that I second every conclusion. Kristian has always been ... special.

 

 

 

[jrhodes]

[HOME AUTOMATION]

I prefer good ol' fashioned REVENGE!

[+Vorticon]

If you are interested in the James Webb Space Telescope, the launch is in less than 3 days. You can sign up as a virtual launch guest here: https://jwst.nasa.gov/content/webbLaunch/countdown.html#

[+Lee Stewart]

37 minutes ago, Vorticon said:

If you are interested in the James Webb Space Telescope, the launch is in less than 3 days. You can sign up as a virtual launch guest here: https://jwst.nasa.gov/content/webbLaunch/countdown.html#

 

We can see it from the beach—though, whether we get up that early to watch it is problematic. 

 

...lee

[GDMike]

Watched Apollo 11 in '69 from the beach after getting Duncan doughnuts first, we had 1 mission..get doughnuts and coffee.  Watched Apollo 12 from the school yard. 

But I've registered to see this giant binocular go up and see how much inflation has risen in just a short sorry crappy year for mankind!  Hopefully those mirrors are programmed correctly this time., But we can always blow it up in a million pieces thank you Russia for figuring that out, soon earth will look like Saturn with a big ring of space garbage surrounding it. Well just call us..."earthRings"

Edited December 21, 2021 by GDMike

[GDMike]

I think these books I bought were in 1983. looking through one..I see this.

eBay has a copy

https://www.ebay.com/itm/234041624077?chn=ps&norover=1&mkevt=1&mkrid=711-117182-37290-0&mkcid=2&itemid=234041624077&targetid=1264870804704&device=m&mktype=&googleloc=9025649&poi=&campaignid=14859008593&mkgroupid=130497710760&rlsatarget=pla-1264870804704&abcId=9300678&merchantid=6296724&gclid=Cj0KCQiAk4aOBhCTARIsAFWFP9HJQ6VGqNSrgdpAl3rVm6l99Lu7G-CBwqKzolDy3Rp7paVOr1ainggaApAVEALw_wcB

Edited December 21, 2021 by GDMike

[+OLD CS1]

@GDMike I credit Fred D'Ignazio with cultivating my interest in computers.  When I was just a wee pup I read his articles in various magazines and his books were in our school library.  Definitely a name I will never forget.

 

I like how the "Data Processing" book is listed as a kid's book (and he describes why.)  Just wonder if we will ever see the "Little Golden Book of C++ Compilers."

[HOME AUTOMATION]

2 hours ago, OLD CS1 said:

Definitely a name I will never forget.

Here's another...

                         

 

                         

 

[wierd_w]

Bohr model is wrong! (Triggered)

 

That should be a 1S(2)2S orbital diagram! Chemistry fail!

 

 

(Lol, but really, that totally IS wrong. Lithium is highly reactive because the outer S orbital desperately wants to not be only half filled/wants to be gone. That is not well displayed with the bohr model.) 

 

Edit: correction, because sleepy

Edited December 21, 2021 by wierd_w

[HOME AUTOMATION]

42 minutes ago, wierd_w said:

Lithium is highly reactive because the outer S orbital desperately wants to not be only half filled/wants to be gone.

Sounds like it would make a good mixer...

 

 

[+Vorticon]

2 hours ago, HOME AUTOMATION said:

Here's another...

                         

 

                         

 

Yes! My absolute favorite technical book in the world.

[+Vorticon]

5 hours ago, OLD CS1 said:

@GDMike I credit Fred D'Ignazio with cultivating my interest in computers.  When I was just a wee pup I read his articles in various magazines and his books were in our school library.  Definitely a name I will never forget. 

 

I like how the "Data Processing" book is listed as a kid's book (and he describes why.)  Just wonder if we will ever see the "Little Golden Book of C++ Compilers."

I loved reading his articles in Compute! magazine. I still remember one where he describes his first encounter with the Eliza program on a CBM Pet ?. Another favorite of mine was Jerry Pournelle's Chaos Manor column in Byte.

[GDMike]

8 hours ago, GDMike said:

 

Here's a pcode

 

https://www.ebay.com/itm/234041624077?chn=ps&norover=1&mkevt=1&mkrid=711-117182-37290-0&mkcid=2&itemid=234041624077&targetid=1264870804704&device=m&mktype=&googleloc=9025649&poi=&campaignid=14859008593&mkgroupid=130497710760&rlsatarget=pla-1264870804704&abcId=9300678&merchantid=6296724&gclid=Cj0KCQiAk4aOBhCTARIsAFWFP9HJQ6VGqNSrgdpAl3rVm6l99Lu7G-CBwqKzolDy3Rp7paVOr1ainggaApAVEALw_wcB

[+Schmitzi]

12 hours ago, GDMike said:

Here's a pcode

 

https://www.ebay.com/itm/234041624077?chn=ps&norover=1&mkevt=1&mkrid=711-117182-37290-0&mkcid=2&itemid=234041624077&targetid=1264870804704&device=m&mktype=&googleloc=9025649&poi=&campaignid=14859008593&mkgroupid=130497710760&rlsatarget=pla-1264870804704&abcId=9300678&merchantid=6296724&gclid=Cj0KCQiAk4aOBhCTARIsAFWFP9HJQ6VGqNSrgdpAl3rVm6l99Lu7G-CBwqKzolDy3Rp7paVOr1ainggaApAVEALw_wcB

 

nope

 

[GDMike]

I can't delete it...my copy paste sucked on my phone that day.. sorry bout that

[+OLD CS1]

[+OLD CS1]

Just what I wanted for Christmas.  At 9:20pm a customer's stand-alone virtual server decided its RAID controller was too hot and took everything off-line.  Hoo-frikken-ray.

[wierd_w]

Why wasnt the virtual server hosted from a redundant iscsi deployment? (Multi-headed SAN host providing iSCSI LUNs with multipath?)

 

Single points of failure for mission critical services ALWAYS bites you in the balls. 

 

You dont actually need a half million dollar failover config for that either: a handful of linux boxes with some FDDI between them and novel use of the md driver would do.

 

(Each linux box has a local raid array, and presents the array as an iscsi lun over the local FDDI SAN, that the "head unit" then aggregates as raid, and doles out external iscsi luns from over the ethernet connection to the virtual server. (Image files mounted and presented as luns) Any of the physical raid controllers going down would just degrade the lun's host, but not take it down. Ideally, you would have two head units, each offering a path to the same virtual raid, and have multipath io and fault tolerance there also. Other than FDDI being a bit pricy, it is totally doable inexpensively.  Rather than have to plan and set that all up, a lot of customers use all-in-one appliances, like those from NetApp, but you dont actually need that to have such tolerance.)

 

 

Did you give them the gift of the best practices sermon at least?

 

[+Vorticon]

Now that is slick!

 

https://www.facebook.com/groups/IBMPCjr/permalink/1592474927753981/

[wierd_w]

I used to own one of those...

 

I entertained getting one, but they are the ULTIMATE red-headed stepchild. Broken IRQ layout. Odd memory map. No true ISA slots. Proprietary connectors...

 

Still, I entertained a silly notion to build a hardware EMS solution on the cartridge ports. (Yes, both ports) The ports DO NOT pass the WE signal, but writes could be enabled other ways.  

 

The path I was envisioning, was to use the second cartridge slot to contain a ROM EMS provider with appropriate header for the ROM Scan routine (meaning EMS would be available at boot time), that contains some "non executable" code. Specifically, an area that is reserved as an address trap, and then a single byte array as a write trap.  Circuitry in the ram portion of the device would latch on a read to the address area, and generate a fake WE signal to drive the SRAM. It would then interpret the trapped address, raise the address lines on the SRAM, then catch the next read, which would designate the byte to write, and raise the needed data lines. It would then turn off the fake WE and wait for the next cycle.

 

I would use similar approaches to select which 64k page to move in or out, etc.

 

As long as programs written for EMS obeyed the interface rules, I could do all that jiggling on the SwitchPage operation, and use a chunk of conventional as the window to hand to calling applications.

 

(They could write on it normally, since it is actual RAM. On SwitchPage, my routine fires which writes the data into the SRAM using the jiggery pokery. The new page is mapped, and a bulk memcopy copies the New page into the low ram window, then the "ready" status is set.  From the calling application's POV, its just really slow EMS.  I could put an ungodly number of pages in though. The intent would be for using EMSDISK on the Jr.)

 

I never got around to actually implementing it though.

 

I dont love the Jr enough to buy one to build a cartridge prototype for.

[+Vorticon]

The PCjr is a quirky machine for sure but still quite usable. Here's mine, upgraded to 640K RAM, V20 CPU, 2 parallel ports, IDE CF-drive, external 3.5" floppy, CD rom and Zip drives, Xircom external network card, PS/2 keyboard adapter with an upgraded IBM keyboard, mouse. Here running GEM

 

 

[+OLD CS1]

4 hours ago, wierd_w said:

Why wasnt the virtual server hosted from a redundant iscsi deployment? (Multi-headed SAN host providing iSCSI LUNs with multipath?)

I guess it has been a long time since you stepped into a very small business.  In these environments you are generally happy they are not running file shares off a desktop and you can get a server through the door in the first place.  I at least provide a good backup solution and replication up to my "cloud."

 

2 hours ago, Vorticon said:

The PCjr is a quirky machine for sure but still quite usable.

My aunt had a PCjr, and it has held in me a place of curiosity ever since.

[wierd_w]

Well, while not a sexy retro computer, I HAVE resumed a very old project I started and then put away.

 

A D.I.Y. luggable computer in a "wood" case.

 

With handle.

 

The project initially started as a silly thing to make, that would use modern modular parts inside, and thus be 100% upgradable. The initial concept was to add some LIPO cells, a charge controller, and an array of tritium illuminators sandwiched to photovoltaic cells, with DC input for more rapid charging. (The Idea(tm) being to make the box fully self recharging given sufficient time), but after crunching some numbers, realized that could never work. (Tritium illuminators do not emit enough useful energy for the task. I would need a pile of them that would be impossibly expensive to source, heavy, and dangerous if somebody kicked it too hard and shattered the ampules. Hence why the project was shelved).

 

The project has been restarted though, because an acquaintance (husband of a coworker) would very much like to play some computer games with me, and I find modern laptops with their garbage intel integrated IGP based graphics, completely unsuitable for the task.

 

I can cobble together some very nice bits of kit inside the large "faux wooden trunk" I picked up at hobby lobby for the original project, and convert it into a fun, fully upgradeable gaming luggable, with A FUCKING REAL GPU inside instead.

 

I intend to slave together several Onn 90W laptop power cables to a shared 3 prong power inlet port, and parallel their outputs to a 300W picoATX.  If I need more juice, I will slave another such picoATX, and more 90W generic chargers inside.

 

With just a laptop spinny disk for swap and an M.2 NGFF SSD, and a GPU on a 90 degree PCIE riser, and some home made supports, along with suitable fan and vent provisions, I should be good.

 

One of the objectives on the bucket list is to shred a cheap USB keyboard apart, then BUILD a keyboard with real switches using its driver board. (Wired the same as the membrane, which will be discarded), along with embedding an LCD panel into the lid of the trunk.  

 

I have gotten the ATX mounting system for the system board installed in it so far. Just need to get some suitable goodies to do the rest of the project as time and budget allow.

 

---------------

 

RE: Small business, no failover

 

Again, you do not need super advanced kit.  I could probably cobble such a beast together for under 5k. Including a second hand server rack and empty rack chasis, the needed FibreChannel interfaces, some suitable system boards (or second hand rack server paperweights) and lots of Linux.

 

After that you just plug each of the multipathed IO ethernet cards into the LAN, turn on the iscsi options in VMWare, and its all cake after that.

 

(You could even cheese on not having hardware raid contollers if this is a small business, by abusing the md driver in linux. All the FibreChannel goodies need is a block device, and md provides one. 

http://linux-iscsi.org/wiki/Fibre_Channel .  It should thus be possible to software raid say, 6 SATA drives (most desktop boards have that many SATA ports) as RAID4 with md, then present the /dev/md0 device to the FC goodies, and present it as an FC connected LUN.  The head unit combines two such connected LUNs as RAID1 for fault tolerance, and doles out iSCSI using mounted image files hosted on that RAID1 device. The expensive bit is the FC HBAs, and switch. The rest you could use cheap PC gear for.)

 

I would have to check current pricing on used crap, but i would also hit up a local industrial salvage yard. They get refuse equipment, including datacenter gear, from the local aerospace operations. (I know for a fact they have some empty racks out there.)  The trick is to not try to jew them down. Ask a price, and if its fair, take it. They usually have no idea what the old datacenter stuff even is anyway. 

 

That plus fleabay, and some elbow grease, I could most probably cobble a mini-SAN together on the cheap. If I were to start a small business, this is what I would likely do, and have the virtual servers running in that same rackmount with proper multipath IO, and if possible, link aggregation (So the ethernet asics installed have fault tolerance). Linux gives a lot of bang for free, just no support. (The intention would be to migrate the LUNs to a REAL storage controller if the company ever grew big enough, which would just be some bulk transfers and some re-garbaging the garbaged parts type arrangements later.)

 

Naturally, I Do Not Say This to people that would want to pervert my good nature into freemium SAN administration.

 

I would however, share that with other monkeys.

 

I totally know what you mean though about small businesses abusing home grade equipment, and then wondering what to do when it breaks spectacularly. You have my condolences.

 

I just happen to be a trained SAN guy who left tech to go wipe asses.  (Due to stress).  If I were to start a business, I at least know best practices for data protection and service reliability.  I would source suitable equipment, and not waste resources on the home appliance trap.

 

It bothers me when best practices are not followed out of some misguided idea that you can cheap out on data protection or service reliability. Single points of failure will ALWAYS bite you in the balls, and it will ALWAYS be costly when it does. having at least a diy version of a proper rig is above and beyond the woefully inadequate "soho" offerings out there, and probably in the same price range. (You just need an actual admin to run it right...)

 

 

 

 

Edited December 25, 2021 by wierd_w