• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

New Outage Schedule

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Sephis

Sephis

Member
​
Joined
Jun 9, 2004
Location
Bruceton, TN
New Outage Schedule
Starting next week our servers will be turned off for several days a week - initially, Tuesday through Thursday. This will allow us to devote more time to scientific development. 24 Jun 2010 21:56:37 UTC ·
Click to expand...

Anybody see that on SETI's front page? :confused: Looks like the Tuesday outages are three days now instead of one. Should make for an interesting Friday as the servers try to handle three days of backlog instead of one.
 
well hopefully this means they are using the extra time to fix the issues that have been causing the weekly (or more frequent) server crashes :)
 
I don't know if it'll fix the crashed or not. The crashes are mostly a result of old, old hardware giving out slowly but surely. :(


The daily stats will be messed up but after a couple of weeks on the new schedule the weekly stats should still be good ... :)
 
Yeah - but the odd thing is they usually have those temporary servers so you know the connection is good but the main servers are down. That's what kept driving me crazy on my main rig until the end of the second day I decided to check another rig - and it was giving me the same error. Glad Sephis posted the new schedule, I've been light on Net usage this week and hadn't checked the SETI boards for awhile ...
 
Looks like they threw on a daily jobs limit as well.

Jeff Cobb said:
The idea behind the jobs per host limit (currently set at 20) is to allow
every host that tries to connect to get some amount of work. This is
at the cost of some of the hosts getting as much work as requested.
This limit is necessary only when we are pegged at our outbound network
limit (but if that limit were raised, we would surely hit some other in
due course).

This limit will be tuned. We've just started working with the three-day
outage protocol (in the interest of doing science, ie, looking for ET).
We're learning how to work with it.
Click to expand...
 
been a while since i crunched any WUs. finally fixed my rig and got her stable.
are you guys still having trouble getting work? i've only been able to get 2 WU's.
i started crunching Einstein for our team but my client hasn't done any SETI work yet :confused:
under projects/resource share both are set to 100(50%). its that right?
 
I have to ask, do you have it set for a 10 day buffer? I just uploaded a bunch of stuff but none of my rigs asked for more work (which is kinda' odd in itself) and I also got an "invalid application description" error and I have NO idea what that's all about. Off to see what's up, now.


Edit:
Well, I think I know what the app description error was all about - I'm using a BIONC client from last century! :p I'll update during the TWT (Tuesday, Wednesday, Thursday) downtime and hopefully that'll go away. I may be behind in my AK apps as well - it has been awhile since I updated those machines.

As far as getting work goes, I don't know what's up with that. There have been odd things going on because Berkeley is limiting WUs - but the limit is MUCH higher than 2 WU's per day. I think they're still trying to balance out the extended outage with the WU downloads so maybe you're hitting part of that issue ... :shrug:
 
Last edited:
i had it set to 7 days. when i suspended Einstein from the projects tab it downloaded 10 more WU's.
it's weird that my client isn't switching between projects :confused:
 
It probably is switching but setting projects to 50/50 sometimes has odd behavior. I've been running SETI and Einstein on my main rig for awhile now and what it seemed to do, long-term, is try to balance the work done on both projects. The problem comes up when, say, Einstein has several days worth of "high priority" WUs to run. BOINC will always run those first. So even if SETI is actually behind in it's "share time" it will still run Einstein. Then when all those "priority" units finally clear out SETI has a huge amount of time built up so that's what it'll try to run. It gets really messy sometimes! LOL!

What I did was set Einstein to run 25% (10% was too much/little). That way, when SETI doesn't have work Einstein will download work and run it - but not a huge chunk of it. So the machine is running another project (Einstein) when there's no SETI work but almost always runs SETI when it's available.

When SETI quits playing musical servers I may go back to 50/50 but for now I'll leave the 75/25 ratio in place.


BTW - Einstein classifies anything up to almost three weeks out as "priority". I've running some stuff due Aug 4 right now and it's running High Priority ... :-/
 
that makes sense. the report deadline for the Einstein work expires this month vs September for the Seti work.
i changed my SETI work to 10 days. Einstein is now set to 0 days with "maintain enough work for 0.25 days".
resource share ratio is 10000:1 or 99.99%:0.01% :p
hopefully after it's done with Einstein it will get back to crunching SETI and only download Einstein when no SETI units are available.
 
b1f2f2.jpg

well i hope this weeks outage is shorter. i goofed and had it set to 0 days.

back to crunching Einstein :comp:
 
Well, it looks like they are working through some things to get the limits back up to reasonable levels. Also, it looks like the VLAR kill apps will soon be outdated. NOTE: This was posted on the 23rd.
23 Jul 2010 15:54:19 UTC
The servers are all up. We are trying to start with a high limit this week to see what happens. The limits currently are:

40 per CPU
320 per GPU

Depending on how it goes, we may set the final day limits to 8x this rather than unlimited. The calculated hope is that this will allow for a 4 day queue filling (3 + 1 for good measure) while not maxing out bandwidth usage. As always, we'll see.

Server software changes this week include the much desired VLAR behavior (not assigning VLAR WUs to GPUs) and a hook in the assimilator for doing RFI filtering at assimilate time (this should reduce the burden on the back end "ntpckr/rfi filter" loop). The VLAR change will not be evident until all previously split work is distributed.
Click to expand...
 
Last edited:
Wow wish I had read up on the three day outages I would have downloaded more units :bang head. I have enough to get me to about Thursday but the rest will be down time while I hope to get some units on Friday.

I will definitely be changing them to 5 days of work units to make sure I have enough.
 
I should have said something about that. :(


I think most of us run the full 10 days, as high as the setting will go. Unless you know you'll be shutting down or something there's no reason not to ... ;)
 
You must log in or register to reply here.

Similar threads

don256us
T32 EOC WIR 20Mar20: 10 mill, 1 mill, 166 and a whole lot of fun.
Replies
3
Views
1K
don256us
don256us
don256us
T32 EOC WIR 20Jun20: 5 Billion and the return to milestones.
Replies
9
Views
842
Slayer2003
Slayer2003
Lochekey
Help with Failing WUs
Replies
17
Views
1K
MaddMutt
MaddMutt
Silver_Pharaoh
Anyone else can't send in WU's?
Replies
17
Views
3K
Silver_Pharaoh
Silver_Pharaoh
rustyfender
failed WU's galore? well that's new
Replies
5
Views
1K
rustyfender
rustyfender
Back