The good news is that recent woes due to lack of workunit disk space have seemingly passed for now. We're still on the very edge of our capacity, but now that we're prioritizing the smaller regular workunits (as opposed to the big Astropulse workunits) we were able to build up a ready-to-send queue and network traffic stabilized overnight.

The less-good news is that we still need to build some indexes on the science database. We're building one now, and it usually takes 12-24 hours. This adds a lot of CPU and disk I/O to the science database server, meaning the splitters can add rows as fast, nor can the assimilators. So the ready-to-send queue drops, and the assimilator queue rises. As an added bonus, when the assimilator queue rises, that means the deleters slow down, which means the available workunit disk space reduces, and we're back to square one again. No big deal as long as people are patient. All the backend services are doing the best they can until the index build finishes, and then we should catch up again.

- Matt

When I run the client it says that Another instance of BOINC is running, when there isn't another instance.

Ok, there would be a lockfile in the boinc directory which can safely deleted when no BOINC client is running. If this lockfile exists at the start of the BOINC client you would get these error message.
You can remove the file "lockfile" in the BOINC directory and start the BOINC client then again.

Nope, same problem even with it gone :(

You may not be seeing the BOINC processes running, depending on how you are querying the process list on *nix.

How are you looking for the BOINC processes?

You should try this command:

ps aux | grep boinc

If you see more than one line in the resulting text, then you most likely have a BOINC process running. (If you only have one line that includes 'grep boinc', it's just finding the grep process.)

In any case, it would help to understand which version of Mandriva, and what version of BOINC you are running, including if you are running it from your "local" directory, or is the BOINC directory somewhere else.

I get this line:

einstein  6256  0.0  0.0   3112   740 pts/1    S+   00:25   0:00 grep --color boinc

Don't chuck the mb yet.

Pull the motherboard out of the case and hook things up with it just sitting on an antistatic pad.

I've seen similar problems that were caused by a motherboard shorting to the case.

HTH

UncleVom

Even when working previously? Unless something metal slid under the MB inside the case and is causing the short, I doubt it would suddenly start shorting on something on its own. But I'll try it anyway....

It's always good to test a mb out of the case when it appears to be the culprit and there is nothing obviously fried, heck it's coming out anyway.

The one I came across with the symptoms was I believe a Asus A7V8* of some variation in a cheap case with large formed standoffs that were causing the short.

It had been working previously but the board must have shifted or the cheap case flexed and the board moved. Once I had determined the problem I added insulating washers, stuffed it back together and it worked fine.

UncleVom

Before the downtime, one of the splitters was off and result creation rate was low. This drained the ready to send queue. Since the servers came back on I guess a lot of hosts still trying to fill their caches, so we are in recovery phase.

When I run the client it says that Another instance of BOINC is running, when there isn't another instance.

Ok, there would be a lockfile in the boinc directory which can safely deleted when no BOINC client is running. If this lockfile exists at the start of the BOINC client you would get these error message.
You can remove the file "lockfile" in the BOINC directory and start the BOINC client then again.

Nope, same problem even with it gone :(

I'm working on a friend's system that has a specific problem: all the lights come on (and stay on) but no video. That is, the DVD-ROM drive access light stays light, so does the CD-RW drive as well as the standard Power LED indicator on the case. No BIOS beeps; no video. All fans spin up, including the CPU fan.

I figured something in the PSU could have blown so I tried a different one that I know is working fine, but no go.

I've never, in all my years, seen something like this, and I don't even know what search phrase to use on Google to see if there's anybody else that has solved this problem.

It would seem to me that it has to be power related being that all the lights remain light but the BIOS doesn't even get a chance to being POST.

Any ideas?

OK, I have been testing some:

When I set:
max_ncpus_pct to 95%
cpu_usage_limit to 100%

The load statistics show no difference to 100%/100%

When I set:
max_ncpus_pct to 100%
cpu_usage_limit to 95%

The load statistics vary between 100% and around 90% (leaving an average window of 5% for other very low priority applications)

Interesting to mention is that I previously set both values to 95%, and when I look back there seems no difference to a 95%/95% setting and the latter 100%/95% setting!

But first you need to have the money to afford a solar system installation up front. Then you recoup the costs through the savings on your power bill. Such an approach might be good over the long term (10 or 20 years or more), but you got to have the up front cash.

Maybe not. Check this out.

Alternately, if you really scrounge, you can find panels for about $2 watt. They won't be new. At that price, a $1000 investment will give you (conservatively) 2.5kwh/day. If you really work on keeping the power draw down, you can probably get a Core 2 Duo P9500 crunching 24/7 off-grid -- and just maybe, you can get two.

Nice, Rented panels, Only fly in the ointment, Sub metered panels and I'm on a sub meter, If I made any excess power I wouldn't get paid for It and therefore somebody gets free power, So for Me It's worthless as I can't get a regular metered account with SCE as the park where I rent the space from is the customer of SCE(Southern California Edison). And yes this sucks, But no one is willing to change this in the legislature. Nothing like being a solar have not.

That is an issue, on a couple of different levels. The net-metering laws could use a little bit of work.

The deal with CitizenRE is a little different: you agree to buy power from them at a fixed price for the duration of the contract -- maybe $0.16/kwh. They put up the solar, and they buy (and sell) the power to SCE.

CitizenRE knows what their costs are, and they know what is happening with utility rates, so it's a straight business proposition for them. Pretty cool.

boinc.exe added to exceptions list, both *.cfg files present.
I emphasize that I can remotely access to BOINC installation on Vista host via BOINC manager, but can't do that via BOINCView.
All another hosts (Win2003 server) on that net are accessible via BOINCView.

I tried browsing their messageboard, but to no avail.

I have all my computers linked in, and I can view their status beautifully. There is a button at the top that says report finished tasks, update projects and get more work when needed. I clicked it expecting it update the computer I selected like it would when I click the update button on the BOINC manager. Nothing happens when I click this.

I selected all computers and clicked in. One reported in, even though about 5 of them had work finished, uploaded and ready to report. When I click the button a message box comes up that says "Retry communication at (computer name or all computers)?"

The "start an update at the selected computer now" button doesn't seem to work either.

Why can't I seem to communicate outwardly with these comps?

EDIT I can seem to run benchmarks on the remote computers.

Hello.
I need remote access via BoincView to BOINC running under Vista.
I managed to provide remote access via BOINC manager, but BOINCView still states "can't connect to host".
What differences between remote access style of BOINC Manager and BOINCView and what needed to do with Vista firewall to allow BOINCView to connect too ?

ADDON: That Vista host doesn't reply on ping requests, maybe it should be changed? If yes, please point how to do this under Vista?

See that the servers are down again. No work. Unable to join other projects because my client is SETI_Enhanced capable ONLY. So why do I see that I'm supposed to be working on an AstroPulse WU in my Task list? Which I don't have! Somebody should be checking on whether a client is capable of crunching AP WU's and not download them to that client. Now, the other person with my supposed WU is going to have to wait for the timeout on me and then wait again for somebody else to crunch that WU.

[quote]Since an otherwise flawless upgrade to 6.2.14, BOINCLOGX can not find my work. I know the folder for BOINCLOG must have changed. Anyone know what it is now? All I can see is the one that BOINCLOG says is not a BOINC folder.
Thanks.

Richard, you need to change the directory..... Since the upgrade to version 6 moved where the application data is housed.

This is what I updated mine too based upon my installation.

C:\Documents and Settings\All Users\Application Data\BOINC

That explains why I could not find it. Hmm. . . Still no good for me. Would you know the change for Vista? The directory you mention does not exist for Vista.

Server status is mostly red or orange.

Yea, it appears the servers are trying to handle the upload spike.

Current Traffic Incoming data is the blue line. Outgoing data are the green bars.

Oh, please, folks. Stop with the artificial anecdotes and the Pollyanna spin doctoring.

Just review the message boards for the history. AP has caused a major disturbance in the system, largely due to apparently unforeseen storage issues and "ghost wu's". I'm all for moving to AP but right now the "try-it-and-see" method advocated below is simply wasting resources for chaotic gain, at best.

Myself, I'm in the "try-it-SLOWLY-AND-METHODICALLY-and-see" camp. But I believe I must be in the minority here. So I shall now get off my soap box, squat under a proper tree, and contemplate my navel.

Yes, AP is seti, or at least the next wave. Everybody probably can understand that, but AP is not the enhanced seti, and so, some may be confused. Plus, AP is not stable and therefore does not deserve to be considered production worthy. (I saw somewhere that even some of the developers aren't sure it is returning correct info-- and so people are hesitant to work on optimized clients for awhile.)

Nobody is getting personal here, about Matt & Co. or anyone else.

What I don't understand, or agree with, is their method. AP was turned on and the system began to puke. You can spin all you want, but the fact is that the server system was generally ok until the transition. Now it isn't. Given that problems were reported right away both by the users (willing and unwilling) and by the inner circle, I would have thought the thing to do would be to turn AP off, or throttle it severely, during periods when nobody was going to be around to patch a problem. This isn't being done. The consequence appears to be a loss of resource (cpu time and/or volunteer clients) that will take time to recapture.

And since neither what I, or anybody else thinks is binding on the project team, one shouldn't get one's panties in a bunch. Perhaps, if we gripe, cajol, and plead enough we may hit on something interesting or previously incorrectly ignored and something positive will come of this message board chatter. The converse isn't helpful at all. Can you imagine waking up every morning and your significant other tells how wonderful the world is with you in it? That would be one definition of Hell.

Here is a useful litmus test: true scientists are very hard to live with for many reasons, not the least of which is that they are trained to analyze and to criticize everything until all imaginable alternatives are exhausted. This behavior generally occurs automatically without consideration of emotion or other human frailties until it is too late.

I've been doing a little window shopping on NewEgg, pending an approval from the Finance Dept (i.e., the g/f), for parts to a new HTPC that is powerful enough for Blu Ray. My main objectives are (in order of importance):

• Powerful enough to run Blu Ray
  
• Cheap enough to satisfy a very cheap budget (as cheap as possible)
  
• Low powered enough to save on heat dissipation (going into a slightly enclosed area with 2 external fans)
  
• Preferably with 8 channel sound
  

I am upgrading from:

• AMD Athlon XP 2600+ 333MHz FSB
  
• ASRock K7VM4 motherboard
  
• 2GB Kingston DDR 333 RAM
  
• ATI Radeon X1650Pro AGP 8x
  
• 120GB Maxtor HD ATA/133
  
• Coolermaster case mATX w/PSU
  
• DVD-ROM and DVD-RW
  

These are the parts I think I will be upgrading to:

• AMD Athlon X2 4850e 2.5GHz 45W 2x512KB $77
  
• Asus M3N78-EMH HDMI w/nVidia 8200 onboard (Blu Ray capable) $89.99
  
• Kingston 2GB (2 x 1GB) DDR2 800 ECC Unbuff (x2, 4GB) $87.98 ($43.99 ea.)
  
• Seagate Barracuda 7200.10 250GB 7200RPM SATA 3.0Gbp/s $59.99
  
• Antec EarthWatts 380W, 80% efficiency, ATX12V v2.0 PSU 20+4 pin connector $29.99
  
• Lite-on Black 4x Blu Ray DVD-ROM SATA Retail $139.99
  

Also buying a beige ATX Midtower case for a dual Athlon MP 2600+ w/4GB DDR 400 that I have lying around for $39.99.

This brings my grand total to $524.93 + $42.07 shipping.

According to my power supply calculator, this system will be drawing about 198W at peak usage. There's no point in buying DDR2 1066 RAM since the board states that only AM2+ CPUs support DDR2 1066, and that seems to be Phenom CPUs only, so I went with the DDR2 800 ECC for added stability.

I was going to shoot for an Intel Core 2 Duo, but they start at $119.99 and 65W output. Plus, Core 2 boards, like the Asus P5E-VM HDMI starts at $129.99. I even tried for a Pentium Dual Core, but they are 65W parts and don't perform as well as the Athlon X2s. Other than better SETI performance, is there any reason to go with a Core 2 + motherboard for approx. $100 more? According to my PSU Calc, the Core 2 setup will run at 185W peak power. Is it worth $100 now for 13W less?

I want to keep the expense as low as possible because I'm saving the rest of our expenditure money for a new Nehalem at the end of the year or the beginning of next year.

Anybody see a better deal that I'm missing? Remember, HTPC+Blu Ray first functionality, SETI crunching secondary. I'm also partial to Seagate hard drives and Asus motherboards, and I like the HDMI output right off the motherboard so I can jack right into my stereo receiver w/full 8 channel sound.

My 1 AP unit is now at 15%, 24 hours down, up to 57h 48m to complete. So its doing about 7% every 12 hours.

Real world estimate : 160 Hours... 24 down... 136 to go :-(

Figured out how to test AP builds standalone yet JD?

Sorry, I haven't even looked at the code yet, busy on another project.

. . . none here - only one box crunchin' though - a single core DOH!

but up ^ down (loads) workin' fine . . .