status : October 27, 2010


log in

Advanced search

Message boards : Technical News : status : October 27, 2010

1 · 2 · 3 · Next
Author Message
Jeff Cobb
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 1 Mar 99
Posts: 110
Credit: 40,367
RAC: 0
United States
Message 1044274 - Posted: 28 Oct 2010, 1:27:39 UTC

Just a quick note. Obviously, jocelyn is up. Mork is recovering.

The purchase orders for both oscar and the new mork went out late today or will go out early tomorrow. It takes a while for these things to work their way through the purchasing pipeline.

We decided to go with HP for these machines. They gave us a very good deal. We are getting two identical (oscar class) machines. I'll post the specs in another note. We hope to have them on hand in about 2 weeks.

At this point, we are discussing what we will do between now and when the new servers are on line.
____________

Profile jrusling
Avatar
Send message
Joined: 8 Sep 02
Posts: 37
Credit: 4,764,889
RAC: 0
United States
Message 1044278 - Posted: 28 Oct 2010, 1:31:20 UTC - in response to Message 1044274.

Thanks for the update. Three weeks sounds pretty good.

____________
http://boincstats.com/signature/-1/user/18390/sig.png

Profile perryjay
Volunteer tester
Avatar
Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 13,699,330
RAC: 12,465
United States
Message 1044279 - Posted: 28 Oct 2010, 1:33:52 UTC - in response to Message 1044274.

Thanks Jeff, welcome back to the land of the living. :-) Copied your message and posted it over in the NC forum to let everyone know. Here's hoping they put a rush on them. Maybe the shipping clerk crunches SETI. :-)
____________


PROUD MEMBER OF Team Starfire World BOINC

KB7RZF
Volunteer tester
Avatar
Send message
Joined: 15 Aug 99
Posts: 9447
Credit: 2,844,355
RAC: 1,849
United States
Message 1044284 - Posted: 28 Oct 2010, 1:54:43 UTC

Awesome new's on the ordering of the new Server's Jeff. Thank you for the update!
____________

Profile Byron Leigh Hatch @ team Carl Sagan
Volunteer tester
Avatar
Send message
Joined: 5 Jul 99
Posts: 2784
Credit: 11,756,274
RAC: 451
Canada
Message 1044285 - Posted: 28 Oct 2010, 1:59:56 UTC

thank you Jeff, for the update


Best Wishes
Byron

Profile Uli
Volunteer tester
Avatar
Send message
Joined: 6 Feb 00
Posts: 9357
Credit: 4,930,799
RAC: 3,645
Germany
Message 1044289 - Posted: 28 Oct 2010, 2:28:04 UTC

Thank you Jeff.
____________
Pluto will always be a planet to me.

Cash Donation Specialist

Niflhum
Send message
Joined: 17 Jan 01
Posts: 2
Credit: 3,537,729
RAC: 0
United States
Message 1044290 - Posted: 28 Oct 2010, 2:28:47 UTC

Great to hear, will be nice to see everything up and running again. My only concern is once everything does get back online, will someone be there baby sitting the servers for the first few hours? I ask because I'm sure they'll get hammered with requests to upload/download work packets by pretty much everyone that takes part in this epic adventure of searching the endless skies and they'll crash and need to be bounced.
____________

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3307
Credit: 16,325,116
RAC: 14,664
Sweden
Message 1044297 - Posted: 28 Oct 2010, 2:58:24 UTC - in response to Message 1044274.

Just a quick note. Obviously, jocelyn is up. Mork is recovering.

At this point, we are discussing what we will do between now and when the new servers are on line.


Take the whole system down until you have installed and thoroughly tested the new servers. If that takes a month or two really doesn't matter. It must be extremely time consuming and not beneficial to the science, just trying to keep the system up, when it is this unstable and unreliable.

It's much better if you spend that time doing some science work, instead of baby sitting unstable machines.

____________

Profile jrusling
Avatar
Send message
Joined: 8 Sep 02
Posts: 37
Credit: 4,764,889
RAC: 0
United States
Message 1044299 - Posted: 28 Oct 2010, 3:13:15 UTC - in response to Message 1044297.

Just a quick note. Obviously, jocelyn is up. Mork is recovering.

At this point, we are discussing what we will do between now and when the new servers are on line.


Take the whole system down until you have installed and thoroughly tested the new servers. If that takes a month or two really doesn't matter. It must be extremely time consuming and not beneficial to the science, just trying to keep the system up, when it is this unstable and unreliable.

It's much better if you spend that time doing some science work, instead of baby sitting unstable machines.


I would prefer them to keep the splitters off line and let things gradually clean up a little.
____________
http://boincstats.com/signature/-1/user/18390/sig.png

Profile Wiggo
Avatar
Send message
Joined: 24 Jan 00
Posts: 5190
Credit: 83,053,983
RAC: 71,699
Australia
Message 1044303 - Posted: 28 Oct 2010, 3:47:51 UTC - in response to Message 1044299.

I would think that the best idea would be to leave the uploads and reporting left on so that all results out in the field can be cleared out which should also get rid of all the ghosts that have built up over the last few months.

Cheers.
____________

cwburch
Send message
Joined: 23 Jul 00
Posts: 50
Credit: 14,511,539
RAC: 33,505
United States
Message 1044305 - Posted: 28 Oct 2010, 3:56:29 UTC

I realize it'll be 2 weeks before the new machines arrive on-site. Then I figure another week to load 'em up, transfer data and check things out before the new stuff actually goes live. MEanwhile, like many others I've got WUs that will expire before that. Also, two of my machines are presently crunching Einstein because they can't get any SETI WUs. Any idea when (and if) we'll be able to do any uploads/downloads prior to the new 'puter's coming online? Or is it so bad that the only realistic timeframe you can offer is, "it'll work when it works"? Thanks.
-Carl

Profile Geek@Play
Volunteer tester
Avatar
Send message
Joined: 31 Jul 01
Posts: 2460
Credit: 83,831,094
RAC: 29,940
United States
Message 1044306 - Posted: 28 Oct 2010, 4:01:07 UTC

I believe the best course would be to allow uploads and reporting only for one day then take the servers offline except for forums and data-driven web pages for the forums.

I have no desire to download new work until after the new servers are installed and fully operational.
____________
Boinc....Boinc....Boinc....Boinc....

gomeyer
Volunteer tester
Send message
Joined: 21 May 99
Posts: 488
Credit: 50,157,953
RAC: 0
United States
Message 1044310 - Posted: 28 Oct 2010, 4:23:44 UTC
Last modified: 28 Oct 2010, 4:28:27 UTC

I have to agree with the general sentiment here. If it can be done safely, turn uploads on and let them report. Let the validator, assimilator, deleter and purger catch up, then shut it down except for the forums until you're comfortable with the new server(s).

[edit] Just please let us know as soon as you know how you intend to handle this so we can plan on whether to shut down for a while, run other projects, clean out dust bunnys or whatever. Thanks! [/edit]

Blake Bonkofsky
Volunteer tester
Avatar
Send message
Joined: 29 Dec 99
Posts: 617
Credit: 46,332,781
RAC: 0
United States
Message 1044316 - Posted: 28 Oct 2010, 4:45:18 UTC - in response to Message 1044310.

Even I have to agree. I would much rather just wait a couple weeks for everything to hopefully be 100% stable again, than have unknown amounts of uptime/downtime in the process. Let everyone upload/report, then let they servers do their jobs. Heck, use the extra downtime to get even more of your science done, if the servers can handle that. No use continually band-aiding things when full replacements are only a couple weeks out.
____________

Profile soft^spirit
Avatar
Send message
Joined: 18 May 99
Posts: 6374
Credit: 28,216,480
RAC: 183
United States
Message 1044322 - Posted: 28 Oct 2010, 5:00:54 UTC

I have to agree with the general sentiment.. low/no load, But.. I do have a request if at all possible.

Can we get a rundown of what you ordered? My inner(okay not so inner) geek is dying to know. "Mork class" times have changed, so what is coming in under the hood?? pleeeease?
____________

Janice

B-Man
Volunteer tester
Send message
Joined: 11 Feb 01
Posts: 253
Credit: 147,366
RAC: 0
United States
Message 1044324 - Posted: 28 Oct 2010, 5:03:16 UTC

Copy and past from another thread

I say turn off WU creation. Leave the upload and reporting online drain out all the in the field WUs. Basically run in cleanup mode for a time then start prep work for the server moves and new servers.

If you think you can turn the project on on low. Set WU limits of 1-10/ cpu core and 4-20 /GPU and limp along until the new servers come in. Set a do not reconnect interval for the server in Boinc to once every 30 minutes. This would allow work to be done but reduce stress on the system.

Addendum: Just leave the forums online if you do decide to do a shutdown. It helps with communication and lets the users stay in touch.

Just going to say good luck with whatever the decision is
____________

BANZAI56
Volunteer tester
Send message
Joined: 17 May 00
Posts: 123
Credit: 33,974,747
RAC: 0
United States
Message 1044332 - Posted: 28 Oct 2010, 5:59:19 UTC

If it matters, looks like I agree with most of the others in this thread.

No new work please!


Just allow a clean up of what's already complete and maybe a few reissues........................but only if safely possible.

SMW
Send message
Joined: 16 May 99
Posts: 21
Credit: 10,624,089
RAC: 3,156
United States
Message 1044334 - Posted: 28 Oct 2010, 6:26:17 UTC - in response to Message 1044274.

We need to let the Berkley guys do WHATEVER is needed to clean it all up. If that means we don't get new units for a while, so be it. We need to do what is best for the project, regardless of what volunteers want. We are here to donate computer time ( and $$$) for the goal, regardless of how that fits into any individual volunteers personal idea.
____________
"It is better to be hated for what you are then to be loved for what you are not"
- Andre Gide (1869-1951)

Berserker
Volunteer tester
Send message
Joined: 2 Jun 99
Posts: 105
Credit: 5,386,463
RAC: 0
United Kingdom
Message 1044340 - Posted: 28 Oct 2010, 7:30:32 UTC

I'm going to suggest a slight variation on the theme.

1) Set the back-off to an hour or more. You want clients phoning home as little as possible to keep the load down.
2) Turn on the uploads, scheduler, validator, file deleter, etc. Keep downloads and splitters off. Leave it like this for a few days.
3) If the DB is holding up at this point, turn on downloads. Leave the splitters off.

Steps 1-2 clear the 'out in the field' work and helps avoid lots of issues with missed deadlines. Step 3 will start to clear the pipes of any other outstanding work without inducing more load.

Obviously only do this if the DB is at least fairly stable. It isn't worth risking another three-day recovery if it's likely the DB will crash again, so better everything is left off if there's more than a little doubt Jocelyn can handle even the limited amount of load.
____________
Stats site - http://www.teamocuk.co.uk - still alive and (just about) kicking.

rob smith
Volunteer moderator
Send message
Joined: 7 Mar 03
Posts: 7674
Credit: 44,853,745
RAC: 75,886
United Kingdom
Message 1044345 - Posted: 28 Oct 2010, 7:58:40 UTC - in response to Message 1044340.

I'd go with that - especially if the resumption of downloads is done on a controlled manner, otherwise the "big users" will load the system up with lots of requests for zillions of WU, while the guy who only needs a couple of WU gets the dreaded "project has no work" message. Logic something like - download a small number of WU to each computer, then back off that computer for at least 12 hours.
____________
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

1 · 2 · 3 · Next

Message boards : Technical News : status : October 27, 2010

Copyright © 2014 University of California