Continuing Progress

Message boards : Technical News : Continuing Progress

To post messages, you must log in.

Previous · 1 · 2

AuthorMessage
Odysseus
Volunteer tester
Avatar

Send message
Joined: 26 Jul 99
Posts: 1807
Credit: 4,913,052
RAC: 19
Canada
Message 1052121 - Posted: 1 Dec 2010, 2:35:13 UTC - in response to Message 1052114.

Well, if it was I In charge, I'd give these guys a call (or a similar vendor)
http://www.racksolutions.com/home/all-products-by-photo/server-racks-cabinets-enclosures.html "Fits any server" and see if they can build me one that works.

They also say “Dell, HP, IBM, Cisco compatible”, which tends to imply that incompatibilities exist, and at any rate what the SSL seems to need is “Fits any collection of half a dozen assorted servers”. ;)

ID: 1052121 · Report as offensive
Profile Bill Walker
Avatar

Send message
Joined: 4 Sep 99
Posts: 3868
Credit: 2,697,159
RAC: 8
Canada
Message 1052124 - Posted: 1 Dec 2010, 2:41:06 UTC - in response to Message 1052121.

Well, if it was I In charge, I'd give these guys a call (or a similar vendor)
http://www.racksolutions.com/home/all-products-by-photo/server-racks-cabinets-enclosures.html "Fits any server" and see if they can build me one that works.

They also say “Dell, HP, IBM, Cisco compatible”, which tends to imply that incompatibilities exist, and at any rate what the SSL seems to need is “Fits any collection of half a dozen assorted servers”. ;)


Unlike many other industries, the computer business has never had much success at generating standards through neutral bodies like ISO or IEC. Maybe because standards are typically on development cycles so much longer than the hardware and software.

personally, I blame Bill Gates.

ID: 1052124 · Report as offensive
John McLeod VII
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 15 Jul 99
Posts: 24806
Credit: 754,585
RAC: 140
United States
Message 1052142 - Posted: 1 Dec 2010, 4:11:32 UTC - in response to Message 1052124.

Well, if it was I In charge, I'd give these guys a call (or a similar vendor)
http://www.racksolutions.com/home/all-products-by-photo/server-racks-cabinets-enclosures.html "Fits any server" and see if they can build me one that works.

They also say “Dell, HP, IBM, Cisco compatible”, which tends to imply that incompatibilities exist, and at any rate what the SSL seems to need is “Fits any collection of half a dozen assorted servers”. ;)


Unlike many other industries, the computer business has never had much success at generating standards through neutral bodies like ISO or IEC. Maybe because standards are typically on development cycles so much longer than the hardware and software.

personally, I blame Bill Gates.

If one standard is good, a dozen standards must be better!


BOINC WIKI

ID: 1052142 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 1052145 - Posted: 1 Dec 2010, 4:20:00 UTC - in response to Message 1052050.

Matt has probably thought of this and has an answer.

Since the work that out in the field on client computers would be minimal when restarting the project perhaps the "resend" option could be turned on at the servers to eliminate the ghost problem.

The resend option only increases server load. So not a good idea.

And as every computer in the universe has no tasks and will be trying to make contact. It would be better to decrease the downloads/processor so that the host gets enough work for a few hours, and set the server enforced back-off to 60 mins and not the default 11 sec. The lower the server load the less likely chance of ghosts.


I would think that having a (now cleaned up) data base that is lean, with no ghost work units waiting to be resent, would be more desirable. I would try to keep the ghosts out of it as they only bloat the data base with no benefits.

Servers are going to be hit hard anyway and that is the precise time when ghosts are created. I would rather receive my resends than have them waiting to time out and get resent to someone else.

Boinc....Boinc....Boinc....Boinc....

ID: 1052145 · Report as offensive
Tom95134

Send message
Joined: 27 Nov 01
Posts: 216
Credit: 3,790,200
RAC: 0
United States
Message 1052147 - Posted: 1 Dec 2010, 4:25:03 UTC

You guys have all the fun!

Nice job. Looking good.

Of course now with everything cleaned up you'll never be able to find anything again.


ID: 1052147 · Report as offensive
uglybiker
Volunteer tester
Avatar

Send message
Joined: 6 Dec 02
Posts: 29
Credit: 9,690,484
RAC: 0
United States
Message 1052167 - Posted: 1 Dec 2010, 7:20:57 UTC - in response to Message 1052066.
Last modified: 1 Dec 2010, 7:21:34 UTC

The main problem with those cheap twist-ties is that the binding is metal, which can cause problems in a high-speed high-density data environment...



Problem solved! :-P

ID: 1052167 · Report as offensive
Profile soft^spirit
Avatar

Send message
Joined: 18 May 99
Posts: 6438
Credit: 31,822,585
RAC: 6,570
United States
Message 1052189 - Posted: 1 Dec 2010, 11:14:19 UTC - in response to Message 1052167.

The main problem with those cheap twist-ties is that the binding is metal, which can cause problems in a high-speed high-density data environment...



Problem solved! :-P


Actually industry standards lean more towards WAXED twine.. But when you get to patch panels they need to be somewhat loose. Which looks messy which makes management think something is wrong. Which is a pain. So wire it how it works and close the door so people do not nag over things that are NOT effecting how it works.

Janice

ID: 1052189 · Report as offensive
Cheopis

Send message
Joined: 17 Sep 00
Posts: 150
Credit: 16,554,617
RAC: 1,778
United States
Message 1052224 - Posted: 1 Dec 2010, 15:07:06 UTC - in response to Message 1051951.

You can get twist ties free on loaves of bread :)

All silliness aside, glad to hear things are moving along nicely.

ID: 1052224 · Report as offensive
Profile Gary CharpentierCrowdfunding Project Donor
Volunteer tester
Avatar

Send message
Joined: 25 Dec 00
Posts: 18630
Credit: 21,420,971
RAC: 20,439
United States
Message 1052258 - Posted: 1 Dec 2010, 17:27:37 UTC - in response to Message 1052189.

The main problem with those cheap twist-ties is that the binding is metal, which can cause problems in a high-speed high-density data environment...



Problem solved! :-P


Actually industry standards lean more towards WAXED twine.. But when you get to patch panels they need to be somewhat loose. Which looks messy which makes management think something is wrong. Which is a pain. So wire it how it works and close the door so people do not nag over things that are NOT effecting how it works.

What may matter more is to have the cable lengths the same so the propagation delays are the same. After all on a multihomed system having different cable lengths for each interface could result in the system picking the one with the shorter length cables for essentially all traffic until that one overloads. That isn't the 50/50 split you want.

ID: 1052258 · Report as offensive
mfbabb2
Avatar

Send message
Joined: 21 Apr 00
Posts: 8
Credit: 132,231
RAC: 0
United States
Message 1052262 - Posted: 1 Dec 2010, 17:48:59 UTC

"The nice thing about Standards is that there are so many to choose from!"

Great work guys! Do whatever you need to do to make life easier in the future -- we can wait -- albeit impatiently in some cases.


ID: 1052262 · Report as offensive
Previous · 1 · 2

Message boards : Technical News : Continuing Progress


 
©2016 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.