"Holy Moly", I just checked my "Stats"" and found Tasks Pending for Validation going all the way back to September.

What's the problem??

I'm concerned since I'm getting very close to my first 1,000,000. mark and there's over 30K in credit "pending".

Just curious ...........

Regards to all,
Jim

---

Clear Skies to all amateur Astronomers out there...

Hi, maybe I was your wingman in November, when I was admitted to hospital and most of the WU's timed out.
A power failure did stop all host, except one which was set to boot after a power down and didn't need a password.
Eventually, if they're not faulty, you'll get credit for them.

Sorry if it was me. I'm sure, you'll get over it, ;^).

Just keep on crunching.

---

Hi There Fred,

Thanks for the info.

'ppreciate the feedback.

Jim

---

Clear Skies to all amateur Astronomers out there...

Jim, you are not the only one. I just sit and wait. My RAC is less than 2,000 and my turnaround is under 3 days.

When I check to see my pending and I see people with over 4,000 RAC and sometimes 5,000 WUs with a turnaround of 40 days, I just get the feeling there is something not right about that.

What really gets me is when after waiting 2 months, you get a No Response timeout from a lot of these and then the wait can start all over again. I guess they are picking and choosing which WU they want to run.

---

SETI@home classic workunits 4,019
SETI@home classic CPU time 34,348 hours

There are a lot of ghost WUs out there as well. I had to take one of my computers off S@H. Before I did that I aborted all the WUs it had. I see the computer is still sitting there with numerous WU with a status of "In Progress". Guaranteed I never received these. I am sure many other WU are in the same state of limbo. It will all wash out in the end though.

Hi Bill,

It's not so much pick and choose as it is ghost work units they don't even know they have on their machines. Many people don't visit the websites so they don't know they have that many workunits assigned to them.

Most of those with such long turn-arounds are running multiple projects with a lot of WUs from each project. Just looking at the BOINC Manager, everything looks ok so they just keep chugging along. All you can do is wait or maybe try sending them a message if they are not anonymous. Sending a message probably won't do any good though since there is little they could do about ghosts other than detach which would cost them any good WUs they have on their machines. Best to just grin and bear it. :-)

---

PROUD MEMBER OF Team Starfire World BOINC

Hi Bill,

It's not so much pick and choose as it is ghost work units they don't even know they have on their machines. Many people don't visit the websites so they don't know they have that many workunits assigned to them.

Most of those with such long turn-arounds are running multiple projects with a lot of WUs from each project. Just looking at the BOINC Manager, everything looks ok so they just keep chugging along. All you can do is wait or maybe try sending them a message if they are not anonymous. Sending a message probably won't do any good though since there is little they could do about ghosts other than detach which would cost them any good WUs they have on their machines. Best to just grin and bear it. :-)

Interesting, never heard of ghost wu before, but then for years I just did my thing without looking at the forums. I have started looking at them just to get a better feel for what is going on lately. Since I do a lot of beta testing, I get to change computers quite often, never the most powerful, always something down the price list somewhere.

---

SETI@home classic workunits 4,019
SETI@home classic CPU time 34,348 hours

Just FYI, ghost WUs are ones the SETI server tries to send to you but they somehow get lost between the server and your machine. I'm not sure what causes it but it usually happens around the times when SETI is having problems.

Another cause of long pendings are people that quit the program without clearing out their caches. I've had some that stay in pending for three or four months because of a combination of those problems.

---

PROUD MEMBER OF Team Starfire World BOINC

<RANT>

Well it seems that ghosts like to haunt my wingmen. I have over 7 days, based on my RAC, of workunits pending with several that can be classified as paired to wingmen who have thousands of workunits that are weeks old being ignored for ones fetched only a few days ago.

The best wingman is the one who hooked up for one day, fetched over 1600 WU and then went dark.

Why am I always paired up with these guys? I had a string of these five or six weeks ago as well. Rarely I fetch more than one per day so it's not like I got the same guy for all of the hanging WUs but sure enough I seem to be batting .500 getting assigned WUs that I have to wait 6 weeks until they're given to someone who actually gets them done in short order.

</RANT>

---

"Life is just nature's way of keeping meat fresh." - The Doctor

<RANT>

Well it seems that ghosts like to haunt my wingmen. I have over 7 days, based on my RAC, of workunits pending with several that can be classified as paired to wingmen who have thousands of workunits that are weeks old being ignored for ones fetched only a few days ago.

The best wingman is the one who hooked up for one day, fetched over 1600 WU and then went dark.

Why am I always paired up with these guys? I had a string of these five or six weeks ago as well. Rarely I fetch more than one per day so it's not like I got the same guy for all of the hanging WUs but sure enough I seem to be batting .500 getting assigned WUs that I have to wait 6 weeks until they're given to someone who actually gets them done in short order.

</RANT>

Well I checked and am not one of your wingmen. It would be a nice feature to be able to select a unit to run if someone asked. It wouldn't bother me to run one out of turn for someone.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

Actually there are multiple parts to the problem.

In some cases it is a user that look and sees the the machine has downloaded a Tons of work and just turns it off... Problem sloved.

In some cases it is a user that has have been advised that a "10 day Cache" solved all the outage problems and has crashed the computer.

In some cases, it is a Server Problem due to Network isues...

It all comes out in the wash... It only takes time...

If I look at the Server Status Page Matt provides some basic information about the "average turn around time" which runs from 60+ hours to over a 100+ hours (for MultiBeam + Cuda). That basically means that anything "pending" should be expected to remain there for about three days. After three days it should be completed (with exception of 10 day caches).

So this goes back to where I have been beat up about advising against larger cache size. The most any user should really set is 3-5 days. The Seti administrators are sensitive to Server Issues and work to get things back to running as quickly as possible. Beyond that you should have a backup project if you are only interested in credit.

For myself (other than one of the Wife's machines), I am set for a connect interval of .25 days and maintain 1 days work. But then I do several projects. My CPU's stay warm.

I would guess that a part is just being "realistic"...
We are working on a problem that will probably not be solved in "Our Lifetime," it does not mean that we should not try. With each of our Completed Results we are one step closer to solving the problem. Yes or No... MY Hope is Yes!
If we were to read all of Matt's Tech News post he is fighting an unwinable hardware battle. Every now and then Matt Wins one. Matt has not quit. He is Patient and Continues.
That is all that can be asked of anyone. To win the battles that each of us can Win... Applause for when "We Win."

Regards, Best Wishes, during the Many World Holidays

---

Please consider a Donation to the Seti Project.

Actually there are multiple parts to the problem.

In some cases it is a user that look and sees the the machine has downloaded a Tons of work and just turns it off... Problem sloved.

In some cases it is a user that has have been advised that a "10 day Cache" solved all the outage problems and has crashed the computer.

In some cases, it is a Server Problem due to Network isues...

It all comes out in the wash... It only takes time...

If I look at the Server Status Page Matt provides some basic information about the "average turn around time" which runs from 60+ hours to over a 100+ hours (for MultiBeam + Cuda). That basically means that anything "pending" should be expected to remain there for about three days. After three days it should be completed (with exception of 10 day caches).

So this goes back to where I have been beat up about advising against larger cache size. The most any user should really set is 3-5 days. The Seti administrators are sensitive to Server Issues and work to get things back to running as quickly as possible. Beyond that you should have a backup project if you are only interested in credit.

For myself (other than one of the Wife's machines), I am set for a connect interval of .25 days and maintain 1 days work. But then I do several projects. My CPU's stay warm.

I would guess that a part is just being "realistic"...
We are working on a problem that will probably not be solved in "Our Lifetime," it does not mean that we should not try. With each of our Completed Results we are one step closer to solving the problem. Yes or No... MY Hope is Yes!
If we were to read all of Matt's Tech News post he is fighting an unwinable hardware battle. Every now and then Matt Wins one. Matt has not quit. He is Patient and Continues.
That is all that can be asked of anyone. To win the battles that each of us can Win... Applause for when "We Win."

Regards, Best Wishes, during the Many World Holidays

If #2 were directed at me my computer is fine, but I did set all my cache back. The large cache does not seem to matter to my machine for some reason. I worked on my machine quite a bit today and edited a lot of pictures and still seems faster than my old Opteron. The ones I lost were due to me screwing up rolling back my video driver a week ago. I will be able to do all these no problem and it will probably look for work in a few days. Finally got all rebranded and the refills and it quit asking for work late last night. I had one error a bit ago other than that all the others were from trying to overclock using the tool in windows instead of the bios. Since I started running at 3.8mhz everything is looking real good and I am watching, if I had any problems I would back off a bit but the guy I got the settings from had lots of help from a whole board of Overclockers and they seemed like they knew what they were doing.
Happy holidays to you and your family.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

We all know "things" Happen, that "We" never expected...

That is part of this particular story.

So how do you explain the "unexpected?"

Happy Holidays

---

Please consider a Donation to the Seti Project.

In the long run, the pending credit doesn't affect your RAC since you do (mostly) eventually get it.

A couple of comments to go along with the excellent answers others have provided:

We're all affected equally by this 'feature'. It's not really worth worrying about.

Seti has, as BOINC projects go, some of the longest deadlines out there. It also has rather stringent validation requirements. Those two things combined can lead to rather long delays for pending credit. The long deadlines are perfectly suited for this type of project, however.

Someone mentioned that some people are 'picking and choosing' which WUs to run. That's possible, but a far more likely scenario is that the BOINC client itself is making those decisions. Although I currently run with a tiny queue (0.04 days), I have in the past run with a 10 day queue on a very fast computer. Many WUs were in the queue from multiple projects. With the long deadlines in the SETI project, the BOINC client had plenty of time to play with and would be constantly executing WUs out of order for some reason. Some WUs would be executed right after they're downloaded, while others would be held until close to their deadline. I found it very annoying, and it certainly was no fun for my wingmen. That's one of the reasons I run with a very short queue these days -- the BOINC scheduler is just weird when it has a big queue and lots of very different projects.

And, finally, there's one more thing to consider: sometimes, computers just up and die. One of my old retired computers was still doing BOINC work booting off a secondary drive after the primary one had failed. Then, for some reason, the second drive stopped working (it was a pretty new drive so I suspect it's a controller or motherboard failure.) Not much can be done about those work units. For that matter, even if I replaced the drive, those work units would still be lost. If a computer dies, a drive crashes, or something else happens to the WU's (including the "ghosts" mentioned above), there's no mechanism for people to tell the server that the WUs are lost.

---

Want to find one of the largest known primes? Try PrimeGrid. Or help cure disease at WCG.

My pending is now over three times my RAC.

Ah well, the system is set up quite redundantly.. so as long as you have processed your wu's correctly you'll get the credit.

I guess not only the project, but also the credit is all about patience and hard work (by our CPU's and the Berkeley crew ;)

PS: my personal highscore, a wu that has been going around since 15 Sep 2009 2:31:45 ;)

---

The SETI@Home Gauntlet 2012 april 16 - 30| info / chat | STATS

Since I posted two of the WU finally completed, only one due to a reissue getting done right away. Now I'm left with 5 (I know that doesn't seem much but it's over 5 times my RAC, down from over 7 times). Of those it looks as if one will get done in about a week, two are on machines that appear defunct now and two are on supercrunchers that appeared to have forgotten about thousands of WUs assigned weeks ago.

And yes, I know, in the long run it doesn't really matter. It's just annoys me when I see rigs grab thousands of WUs, because they are some multi-core with multiple GPU rigs with large queues, and then misplace them due to an instability of the machine or some other problem.

It bothers me more when people running supercrunchers complain about people like me, running a 10 year old cruncher, claiming people like me, with four or five WU in progress, are hogging all of "their" WUs.

In the end, as long as you're getting WUs done in less time than your queue length, it doesn't really matter IMO.

---

"Life is just nature's way of keeping meat fresh." - The Doctor

My pending is now over three times my RAC.

Mine is over 4x. What would be considered "normal"?

Depends on how many WU's you crunch per 24 hrs.
My Pending goes between 300 and 375 thousands but my 2 rigs do about 700 WU'S per 24 hrs.

My pending is about 5x at the moment
Dave

---

Ok, I have a pending total of 12,946.4 current RAC is 1909.49 so pending is 6.8 times my RAC. My oldest pending started life on 18 September. It has gone through 1 detach, 2 inconclusive, 1 timeout, and will timeout again tomorrow. :-(

It's an angle range 0.42 claiming 112.23 credits so it's not a troublesome WU, just luck of the draw on the wingmen. With my luck the next victim will have a 30 or 40 day turnaround!

---

PROUD MEMBER OF Team Starfire World BOINC